fix(security): embed the Raycast signature secret via resolvePublicCred (HR#11)

The secret-scan ratchet only ran once the earlier Fast Quality Gates steps
stopped failing, and it lands at 1 finding vs baseline 0.

The finding is open-sse/services/raycast.ts:19 —
RAYCAST_DEFAULT_SIG_SECRET, a 64-hex request-signature secret that #8895
committed as a bare string literal. It is genuinely public (community-extracted
from the Raycast macOS client; the SAME value ships to every install, it is not
a per-user credential), which is exactly the category Hard Rule #11 governs:
public upstream credentials MUST go through resolvePublicCred()
(open-sse/utils/publicCreds.ts), never a literal — see
docs/security/PUBLIC_CREDS.md.

So the fix is the mandated pattern, not a .gitleaks.toml allowlist entry: added
`raycast_sig_secret` to EMBEDDED_DEFAULTS as the XOR-masked byte sequence and
resolved it with the existing RAYCAST_SIG_SECRET env override. The
providerSpecificData.sigSecret override is untouched. Verified the decoded value
is byte-identical to the literal it replaces.

check:secrets secretFindings 1 -> 0. check:public-creds exit 0.
publicCreds 12/12, raycast-auth 6/6, raycast-local-extract 1/1,
trae-publiccred 3/3. typecheck:core exit 0.

Refs #9298

.env.example

@@ -345,11 +345,8 @@ ALLOW_API_KEY_REVEAL=false
 # OMNIROUTE_CHAT_HEAVY_TOOL_COUNT=64
 # Conservative string-size token estimate that classifies a request as heavyweight. Default 32000.
 # OMNIROUTE_CHAT_HEAVY_ESTIMATED_TOKENS=32000
-# Optional opt-in hard message-count cap; excess receives compact-required 413 before
-# compression can run. Unset/0 (the default) means no history cap: heap growth is bounded
-# by OMNIROUTE_CHAT_MAX_HEAVY_IN_FLIGHT and the heap-pressure shed instead. Set a positive
-# value only on memory-constrained deployments that need a hard ceiling.
-# OMNIROUTE_CHAT_HARD_MAX_MESSAGES=0
+# Hard message-count cap; excess receives compact-required 413. Default 800.
+# OMNIROUTE_CHAT_HARD_MAX_MESSAGES=800
 
 # Hard cap (bytes) for a non-streaming upstream response buffered fully into memory
 # (#5152). Past this the upstream reader is cancelled and the request fails fast
@@ -797,16 +794,6 @@ PROVIDER_LIMITS_SYNC_SPACING_MS=1500
 # Disable the proactive recovery scheduler entirely (default: false).
 # OMNIROUTE_DISABLE_CONNECTION_RECOVERY=false
 
-# Proactive Claude warmup scheduler (#8848): fires a trivial request to opted-in
-# OAuth connections on a cron schedule (America/Los_Angeles) so accounts do not
-# hit the 5-hour sliding window cold. Off by default — set ENABLED=1 and flip
-# per-connection flags in settings.claudeWarmup.connections to activate.
-# Used by: src/lib/warmupScheduler.ts.
-# OMNIROUTE_WARMUP_ENABLED=false
-# OMNIROUTE_WARMUP_CRON="0 7 * * *"
-# OMNIROUTE_WARMUP_CONCURRENCY=3
-# OMNIROUTE_WARMUP_MODEL=
-
 # Background job interval for budget reset checks (ms).  Default: 600000 (10m).
 # Used by: src/lib/jobs/budgetResetJob.ts.  Floor: 10000.
 #OMNIROUTE_BUDGET_RESET_JOB_INTERVAL_MS=600000
@@ -1540,15 +1527,6 @@ APP_LOG_TO_FILE=true
 # ═══════════════════════════════════════════════════════════════════════════════
 # 19. MODEL SYNC (Dev)
 # ═══════════════════════════════════════════════════════════════════════════════
-# Enable the models.dev capability sync.  Default: false (opt-in only).
-# Also settable from Dashboard > Settings > AI. This variable wins over that
-# setting whenever it is set to anything non-empty, in either direction, so a
-# deployment can pin the sync on or off without depending on database state
-# surviving a rebuild. Leave it unset to let the dashboard toggle decide.
-# On: 1, true, yes or on (any casing). Any other value is off.
-# Used by: src/lib/modelsDevSync.ts
-# MODELS_DEV_SYNC_ENABLED=false
-
 # Development-time model catalog sync interval in seconds.
 # Used by: src/lib/modelsDevSync.ts
 # Default: 86400 (24 hours)
@@ -2492,10 +2470,10 @@ QUOTA_STORE_DRIVER=sqlite              # sqlite | redis
 # ═══════════════════════════════════════════════════════════════════════════════
 # Optional add-on (feature flag RADAR_ENABLED, default off — see feature flag
 # settings, not an env var) that overlays a signed, freshly-curated free-model
-# catalog on top of the release baseline. All four variables below are optional
-# and only needed to point the client at a self-hosted/forked feed or
-# supporter-key flow instead of the default OmniRoute Radar service. Used by:
-# src/lib/radar/sync.ts, src/lib/radar/pinnedKeys.ts, src/lib/radar/links.ts.
+# catalog on top of the release baseline. Both variables below are optional and
+# only needed to point the client at a self-hosted/forked feed instead of the
+# default OmniRoute Radar feed. Used by: src/lib/radar/sync.ts,
+# src/lib/radar/pinnedKeys.ts.
 
 # Base URL of the Radar feed service. Overrides the built-in default so forks
 # and self-hosters can point at their own signed feed.
@@ -2505,12 +2483,3 @@ QUOTA_STORE_DRIVER=sqlite              # sqlite | redis
 # signature, replacing the pinned default key. Required when self-hosting a
 # feed signed with a different key pair.
 # RADAR_FEED_PUBKEY=
-
-# URL the dashboard's "I'm a contributor" button opens (GitHub OAuth
-# supporter-key claim flow). No pricing/value lives in this repo — only the
-# link.
-# RADAR_CONTRIBUTOR_CLAIM_URL=https://radar.omniroute.online/auth/github
-
-# URL the dashboard's "Support the project" button opens (payment/plans
-# page). No pricing/value lives in this repo — only the link.
-# RADAR_SUPPORTER_PLANS_URL=https://radar.omniroute.online/planos

.github/workflows/codeql.yml

@@ -22,10 +22,10 @@ jobs:
       - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           persist-credentials: false
-      - uses: github/codeql-action/init@f205ea1c3313d32999d8d6a48b4f6530d4437b38 # v4.37.4
+      - uses: github/codeql-action/init@e4fba868fa4b1b91e1fdab776edc8cfbe6e9fb81 # v4.37.3
         with:
           languages: javascript-typescript
           queries: security-extended
-      - uses: github/codeql-action/analyze@f205ea1c3313d32999d8d6a48b4f6530d4437b38 # v4.37.4
+      - uses: github/codeql-action/analyze@e4fba868fa4b1b91e1fdab776edc8cfbe6e9fb81 # v4.37.3
         with:
           category: "/language:javascript-typescript"

.github/workflows/docker-publish.yml

@@ -137,13 +137,13 @@ jobs:
         uses: docker/setup-buildx-action@v4
 
       - name: Login to Docker Hub
-        uses: docker/login-action@v4.6.0
+        uses: docker/login-action@v4.5.2
         with:
           username: ${{ secrets.DOCKERHUB_USERNAME }}
           password: ${{ secrets.DOCKERHUB_TOKEN }}
 
       - name: Login to GitHub Container Registry
-        uses: docker/login-action@v4.6.0
+        uses: docker/login-action@v4.5.2
         with:
           registry: ghcr.io
           username: ${{ github.actor }}
@@ -237,13 +237,13 @@ jobs:
         uses: docker/setup-buildx-action@v4
 
       - name: Login to Docker Hub
-        uses: docker/login-action@v4.6.0
+        uses: docker/login-action@v4.5.2
         with:
           username: ${{ secrets.DOCKERHUB_USERNAME }}
           password: ${{ secrets.DOCKERHUB_TOKEN }}
 
       - name: Login to GitHub Container Registry
-        uses: docker/login-action@v4.6.0
+        uses: docker/login-action@v4.5.2
         with:
           registry: ghcr.io
           username: ${{ github.actor }}
@@ -372,7 +372,7 @@ jobs:
       - name: Upload Trivy SARIF to Security tab
         if: needs.prepare.outputs.version != 'main'
         continue-on-error: true
-        uses: github/codeql-action/upload-sarif@v4.37.4
+        uses: github/codeql-action/upload-sarif@v4.37.3
         with:
           sarif_file: trivy-results.sarif
           category: trivy-image

.github/workflows/quality.yml

@@ -271,10 +271,6 @@ jobs:
       # covered by typecheck:core's curated allowlist. See check-dashboard-typecheck.mjs.
       - name: Typecheck (dashboard)
         run: npm run check:dashboard-typecheck
-      # #8781: open-sse workspace typecheck gate — the workspace imports @/ which
-      # escapes to src/ via undeclared path aliases. See check-open-sse-typecheck.mjs.
-      - name: Typecheck (open-sse)
-        run: npm run check:open-sse-typecheck
       # WS4.2 (v3.8.49 plan): TypeScript 7 native-compiler SHADOW — advisory only.
       # TS7 went GA 2026-07-08 with 8-12x type-check speedups; its Compiler API only
       # arrives in 7.1, so typescript-eslint / type-coverage / Stryker stay on 6.x

@omniroute/opencode-plugin/README.md

@@ -30,7 +30,7 @@ omniroute setup opencode --auth
 # 3. Restart OpenCode — /models lists the full live catalog
 ```
 
-The `--auth` flag runs `opencode auth login --provider opencode-omniroute` automatically.
+The `--auth` flag runs `opencode auth login --provider omniroute` automatically.
 Use `--base-url` to point at a non-default OmniRoute address:
 
 ```sh
@@ -84,7 +84,7 @@ Peer dep: `@opencode-ai/plugin` (managed by your OpenCode install).
 ```
 
 ```sh
-opencode auth login --provider opencode-omniroute
+opencode auth login --provider omniroute
 # prompts for the OmniRoute API key, writes to ~/.local/share/opencode/auth.json
 ```
 
@@ -164,8 +164,8 @@ Then in `~/.config/opencode/opencode.json` reference each directory by absolute
 Paths are relative to `~/.config/opencode/`. Each entry now resolves to a distinct module file, so OC loads them as two separate plugin instances. Authenticate each:
 
 ```sh
-opencode auth login --provider opencode-omniroute
-opencode auth login --provider opencode-omniroute-preprod
+opencode auth login --provider omniroute
+opencode auth login --provider omniroute-preprod
 ```
 
 Each entry gets its own provider id, its own model picker entry, its own slot in `auth.json`, and its own TTL cache. Closures are isolated per plugin instance — no cross-talk.

_tasks

@@ -1 +0,0 @@
-/home/diegosouzapw/dev/proxys/OmniRoute/_tasks

bin/cli/commands/setup-open-code.mjs

@@ -218,26 +218,6 @@ function registerPluginInOpenCodeConfig({
  * a clear "could not run opencode" message instead of a hard import
  * failure.
  */
-/**
- * Resolve the provider id used for `opencode auth login --provider <id>`.
- *
- * The bundled @omniroute/opencode-plugin registers its provider under
- * `opencode-<id>` (the `opencode-` prefix is required by OpenCode >=1.17.8's
- * native-adapter gate). The auth login command must use the prefixed form
- * because OpenCode resolves `--provider <id>` against the provider id the
- * plugin actually registered.
- *
- * Idempotent: if the id already starts with `opencode-`, it passes through
- * unchanged. This protects users who manually worked around the bug with
- * `--provider opencode-omniroute`.
- *
- * @param {string} providerId
- * @returns {string}
- */
-export function resolveOpenCodeAuthProviderId(providerId) {
-  return providerId.startsWith("opencode-") ? providerId : `opencode-${providerId}`;
-}
-
 /**
  * Pure resolver for the `opencode auth login` spawn descriptor. Extracted so the
  * platform-branching logic is unit-testable without mocking child_process or
@@ -251,23 +231,21 @@ export function resolveOpenCodeAuthProviderId(providerId) {
  */
 export function resolveOpenCodeAuthSpawn(providerId, platform = process.platform) {
   const isWin = platform === "win32";
-  const authProviderId = resolveOpenCodeAuthProviderId(providerId);
   return {
     command: isWin ? "opencode.cmd" : "opencode",
-    args: ["auth", "login", "--provider", authProviderId],
+    args: ["auth", "login", "--provider", providerId],
     options: { stdio: "inherit", shell: isWin },
   };
 }
 
 export function runOpenCodeAuth(providerId) {
-  const authProviderId = resolveOpenCodeAuthProviderId(providerId);
   const { command, args, options } = resolveOpenCodeAuthSpawn(providerId);
   const res = spawnSync(command, args, options);
   if (res.error) {
     // ENOENT = opencode is not on PATH
     if (res.error.code === "ENOENT") {
       printInfo(
-        `opencode CLI not found on PATH. Run \`opencode auth login --provider ${authProviderId}\` manually after installing OpenCode.`
+        `opencode CLI not found on PATH. Run \`opencode auth login --provider ${providerId}\` manually after installing OpenCode.`
       );
       return 1;
     }
@@ -365,8 +343,7 @@ export async function runSetupOpenCodeCommand(opts = {}) {
   if (wantsAuth) {
     if (nonInteractive) {
       printInfo(`Skipping \`opencode auth login\` (non-interactive mode).`);
-      const authProviderId = resolveOpenCodeAuthProviderId(providerId);
-      printInfo(`Run manually: opencode auth login --provider ${authProviderId}`);
+      printInfo(`Run manually: opencode auth login --provider ${providerId}`);
     } else {
       printHeading("Authenticating with OpenCode");
       const authExit = runOpenCodeAuth(providerId);
@@ -375,9 +352,8 @@ export async function runSetupOpenCodeCommand(opts = {}) {
       }
     }
   } else {
-    const authProviderId = resolveOpenCodeAuthProviderId(providerId);
     printInfo(
-      `Next step: opencode auth login --provider ${authProviderId}   (pass --auth to do this automatically)`
+      `Next step: opencode auth login --provider ${providerId}   (pass --auth to do this automatically)`
     );
   }
 

bin/cli/runtime/trayRuntime.ts

@@ -17,7 +17,7 @@ export const SYSTRAY_VERSION = "2.1.4";
 const SYSTRAY_SPEC = `${SYSTRAY_PACKAGE}@${SYSTRAY_VERSION}`;
 
 export function resolveSystrayBinName(platform: NodeJS.Platform): string | null {
-  if (platform === "win32") return "tray_windows_release.exe";
+  if (platform === "win32") return null;
   if (platform === "darwin") return "tray_darwin_release";
   return "tray_linux_release";
 }
@@ -45,6 +45,7 @@ export function chmodSystrayBinAt(runtimeRoot: string, platform: NodeJS.Platform
 }
 
 export async function loadSystray(): Promise<(new (...args: unknown[]) => unknown) | null> {
+  if (process.platform === "win32") return null; // Windows uses tray.ps1 instead
   ensureRuntimeDir();
   if (!isInstalled()) {
     try {

bin/cli/sqlite.mjs

@@ -130,7 +130,7 @@ async function openSqliteDatabase(dbPath, options = {}) {
   try {
     return new loaded.Database(dbPath, options);
   } catch (error) {
-    return openWithSyncDriverFallback(dbPath, options, error);
+    throw createSqliteNativeError(error);
   }
 }
 

bin/cli/tray/autostart.mjs

@@ -167,10 +167,6 @@ export function getAutostartStatus() {
       linger: tryReadLingerEnabled(),
     };
   }
-  if (process.platform === "win32") {
-    const winMechanism = isAutostartEnabled() ? "vbs-startup" : null;
-    return { enabled: isAutostartEnabled(), mechanism: winMechanism };
-  }
   return { enabled: isAutostartEnabled(), mechanism: null };
 }
 

bin/cli/tray/index.mjs

@@ -1,4 +1,5 @@
 import { isTraySupported, initSystrayUnix, killSystrayUnix } from "./traySystray.mjs";
+import { initWinTray, killWinTray } from "./trayWindows.mjs";
 
 let active = null;
 
@@ -9,17 +10,15 @@ export async function initTray({ port, onQuit, onOpenDashboard, onShowLogs }) {
   const ctx = { port, onQuit, onOpenDashboard, onShowLogs };
   // initSystrayUnix is async: it lazily installs/loads systray2 from the runtime
   // dir (trayRuntime.ts) rather than from node_modules. (#4605)
-  // Use systray2 on all platforms including Windows — the tarball ships
-  // tray_windows_release.exe, avoiding the Norton/AVG IDP.HELU.PSE85 heuristic
-  // that fires on temp-dir PowerShell scripts. (#8609)
-  active = await initSystrayUnix(ctx);
+  active = process.platform === "win32" ? initWinTray(ctx) : await initSystrayUnix(ctx);
   return active;
 }
 
 export function killTray() {
   if (!active) return;
   try {
-    killSystrayUnix(active);
+    if (process.platform === "win32") killWinTray(active);
+    else killSystrayUnix(active);
   } catch {}
   active = null;
 }

bin/mcp-server.mjs

@@ -3,7 +3,7 @@
 import { spawn } from "node:child_process";
 import { existsSync } from "node:fs";
 import { dirname, join } from "node:path";
-import { fileURLToPath, pathToFileURL } from "node:url";
+import { fileURLToPath } from "node:url";
 
 const __filename = fileURLToPath(import.meta.url);
 const __dirname = dirname(__filename);
@@ -43,15 +43,7 @@ export async function startMcpCli(rootDir = ROOT) {
   }
 
   // `tsx` loader is only required for local `.ts` fallback; JS entry works without it.
-  const tsxLoaderArgs = mcpEntry.endsWith(".ts") ? ["--import", "tsx"] : [];
-  // Preload the stdout/stderr console guard before mcpEntry's own module graph evaluates —
-  // DB init (a side effect of createMcpServer()'s tool registration) logs via plain
-  // console.log, and by the time any code inside mcpEntry itself could redirect it, that
-  // module's own (hoisted) imports have already run. Loading the guard first, in a separate
-  // module, is the only point early enough to guarantee it never leaks into the JSON-RPC
-  // stream on stdout.
-  const consoleGuard = pathToFileURL(join(__dirname, "mcpStdioConsoleGuard.mjs")).href;
-  const loaderArgs = ["--import", consoleGuard, ...tsxLoaderArgs];
+  const loaderArgs = mcpEntry.endsWith(".ts") ? ["--import", "tsx"] : [];
 
   await new Promise((resolve, reject) => {
     const child = spawn(process.execPath, [...loaderArgs, mcpEntry], {

bin/mcpStdioConsoleGuard.mjs

@@ -1,16 +0,0 @@
-// Preloaded (via `node --import`) before open-sse/mcp-server/server.ts and its entire
-// import graph evaluate. The stdio MCP transport uses stdout exclusively for JSON-RPC
-// messages, but DB init (getDbInstance(), triggered as a side effect of evaluating the
-// server's module graph — e.g. tool registration reading compression settings) logs via
-// plain console.log. A redirect placed *inside* server.ts (even at the top of its first
-// executed function) is too late: static imports are hoisted and fully evaluated before
-// any of that function's own code runs, so earlier console.log calls during import-time
-// side effects already escaped to the real stdout by then. Redirecting here, in a module
-// that loads before server.ts is even requested, is the only point early enough to
-// guarantee no startup output leaks into the JSON-RPC stream and corrupts it client-side
-// (e.g. Claude Desktop: "Unexpected token 'D', \"[DB] Changi\"... is not valid JSON").
-import { Console } from "node:console";
-
-const stderrConsole = new Console({ stdout: process.stderr, stderr: process.stderr });
-console.log = stderrConsole.log.bind(stderrConsole);
-console.warn = stderrConsole.warn.bind(stderrConsole);

bin/omniroute.mjs

@@ -43,19 +43,6 @@ if (isVersionFastPath(process.argv)) {
   process.exit(0);
 }
 
-// MCP stdio transport uses stdout exclusively for JSON-RPC messages. Redirect
-// console.log/warn to stderr before anything else runs — including the tsx/esm and
-// polyfill imports below, since those (and their transitive module graphs, e.g. DB
-// init) can themselves log during evaluation. Redirecting after those imports let
-// early output leak straight into the JSON-RPC stream and corrupt it client-side
-// (e.g. Claude Desktop: "Unexpected token 'D', \"[DB] Changi\"... is not valid JSON").
-if (process.argv.includes("--mcp")) {
-  const { Console } = await import("node:console");
-  const stderrConsole = new Console({ stdout: process.stderr, stderr: process.stderr });
-  console.log = stderrConsole.log.bind(stderrConsole);
-  console.warn = stderrConsole.warn.bind(stderrConsole);
-}
-
 // Register tsx so dynamic imports of .ts source files (referenced as .js per
 // TypeScript conventions) resolve correctly. The build never emits .js for
 // src/lib/cli-helper/, so tsx handles the .ts → .js resolution at runtime.
@@ -71,6 +58,16 @@ await import("../open-sse/utils/setupPolyfill.ts");
 const { registerAliasResolver } = await import("./aliasResolver.mjs");
 await registerAliasResolver(ROOT);
 
+// MCP stdio transport uses stdout exclusively for JSON-RPC messages.
+// Redirect console.log/warn to stderr early (before loadEnvFile and DB init)
+// so no startup output corrupts the protocol.
+if (process.argv.includes("--mcp")) {
+  const { Console } = await import("node:console");
+  const stderrConsole = new Console({ stdout: process.stderr, stderr: process.stderr });
+  console.log = stderrConsole.log.bind(stderrConsole);
+  console.warn = stderrConsole.warn.bind(stderrConsole);
+}
+
 // Electron persists secrets (JWT_SECRET, API_KEY_SECRET, STORAGE_ENCRYPTION_KEY) to
 // `<DATA_DIR>/server.env` (electron/main.js), never `.env`. Migrating an existing
 // install (storage.sqlite + server.env) to the CLI left those secrets undiscoverable —

bin/restore-policies.sh

@@ -39,8 +39,7 @@ snap="$(ops_find_snapshot "$ID")"
 
 # Policy definition tables present in BOTH the snapshot and the live DB. GLOB
 # keeps `_` literal; we drop usage counters / logs so accounting isn't rewound.
-tables=()
-while IFS= read -r t; do tables+=("$t"); done < <(
+readarray -t tables < <(
   sqlite3 "$snap/storage.sqlite" \
     "SELECT name FROM sqlite_master WHERE type='table' AND name GLOB 'api_key*' \
        AND name NOT GLOB '*counter*' AND name NOT GLOB '*_log*' ORDER BY name;"

changelog.d/features/6736-response-content-encoding.md

@@ -1 +0,0 @@
-- **feat(api):** add response content encoding verification — confirms Next.js compress:true and documents stripStaleForwardingHeaders behavior ([#6736](https://github.com/diegosouzapw/OmniRoute/issues/6736))

changelog.d/features/6752-plugins-marketplace-install-api.md

@@ -1 +0,0 @@
-- **feat(api):** add plugins marketplace install endpoint with checksum verification ([#6752](https://github.com/diegosouzapw/OmniRoute/issues/6752))

changelog.d/features/7679-chatgpt-web-thinking-tool-emulation.md

@@ -1 +0,0 @@
-- **feat(chatgpt-web):** harden prompt-emulated tool contract for thinking models (#7679 — thanks @horacecar)

changelog.d/features/9248-video-url-passthrough.md

@@ -1 +0,0 @@
-- **feat(providers):** make video_url passthrough configurable per provider/model via compat override ([#9248](https://github.com/diegosouzapw/OmniRoute/issues/9248)) — thanks @HellFiveOsborn

changelog.d/features/9270-provider-api-key-links.md

@@ -1 +0,0 @@
-- **feat(dashboard):** render a conditional "Get API key" link on the provider detail page, surfaced from the existing `notice.apiKeyUrl` / `notice.signupUrl` catalog metadata (e.g. `pioneer`, `jina`, `together`). The link opens in a new tab and is hidden when neither URL is present, so existing providers are unaffected. Tracks the notice field in `ProviderCatalogMetadata` ([#9270](https://github.com/diegosouzapw/OmniRoute/pull/9270))

changelog.d/features/9284-json-cookie-input.md

@@ -1 +0,0 @@
-- **feat(providers):** accept JSON cookie objects in normalizeSessionCookieHeader (#9284 — thanks @AIB1TAL0S)

changelog.d/features/9318-opencode-zen-reasoning-effort.md

@@ -1 +0,0 @@
-- **feat(providers):** support max reasoning effort for opencode-zen DeepSeek models (#9318)

changelog.d/features/9570-plugin-context-headers.md

@@ -1 +0,0 @@
-- feat(plugins): expose client request headers in plugin onRequest/onResponse context (#9570)

changelog.d/features/9579-soniox-audio-provider.md

@@ -1 +0,0 @@
-- **feat(audio):** Soniox STT + TTS provider (`sx`) — async speech-to-text (`stt-async-v5`, `stt-async-v4`) and real-time text-to-speech (`tts-rt-v1`) ([#9579](https://github.com/diegosouzapw/OmniRoute/pull/9579))

changelog.d/fixes/8577-fix.plan.md

@@ -1,2 +0,0 @@
-- fix(tests): make machineId tests macOS-compatible by stubbing ioreg in test helper (#8577)
-- fix(scripts): replace bash 4+ readarray with compatible while-read loop in restore-policies.sh (#8577)

changelog.d/fixes/8609-fix.plan.md

@@ -1 +0,0 @@
-- fix(cli): enable systray2 on Windows for Norton-friendly tray (#8609)

changelog.d/fixes/8681-fix.plan.md

@@ -1 +0,0 @@
-- fix(providers): gate premium opencode-zen/opencode-go models behind an API key (#8681)

changelog.d/fixes/8739-yuanbao-sse-parser.md

@@ -1 +0,0 @@
-- fix(yuanbao-web): accept `content` field in SSE text events (upstream format change) (#8739)

changelog.d/fixes/8781-fix.plan.md

@@ -1 +0,0 @@
-- fix(build): remove misleading open-sse/package.json facade and add workspace typecheck gate (#8781)

changelog.d/fixes/8813-chatgpt-sentinel.md

@@ -1 +0,0 @@
-- fix(errorClassifier): classify ChatGPT Web SENTINEL_BLOCKED 403 as terminal FORBIDDEN, enabling proper combo fallback (#8813)

changelog.d/fixes/8826-fix.plan.md

@@ -1 +0,0 @@
-- fix(cli): fall back to node:sqlite when better-sqlite3 constructor throws at runtime (#8826)

changelog.d/fixes/8830-fix.plan.md

@@ -1 +0,0 @@
-- fix(opencode): prefix provider id with "opencode-" for auth login command (#8830)

changelog.d/fixes/8841-fix.plan.md

@@ -1 +0,0 @@
-- fix(opencode-zen): add current free-tier models to registry to enable combo context pre-filtering (#8841)

changelog.d/fixes/8869-opencode-complete-model-limits.md

@@ -1 +0,0 @@
-- **fix(opencode):** generate schema-complete model limits so OpenCode accepts catalog entries without an explicit output cap ([#8869](https://github.com/diegosouzapw/OmniRoute/pull/8869)) — thanks @xiaoyaner0201

changelog.d/fixes/8876-codex-responses-wire-default.md

@@ -1 +0,0 @@
-- **fix(cli):** default omitted Codex CLI wire API settings to Responses and clear stale Chat state after reset ([#8876](https://github.com/diegosouzapw/OmniRoute/pull/8876)) — thanks @xiaoyaner0201

changelog.d/fixes/8883-proxy-credential-autofill.md

@@ -1 +0,0 @@
-- **fix(proxy):** isolate new proxy credential fields from browser and password-manager autofill after form reset ([#8883](https://github.com/diegosouzapw/OmniRoute/pull/8883)) — thanks @xiaoyaner0201

changelog.d/fixes/8921-codebuddy-cn-dual-auth-actions.md

@@ -1 +0,0 @@
-- **fix(providers):** expose both OAuth Connect and manual API-key actions for dual-auth providers such as CodeBuddy CN ([#8921](https://github.com/diegosouzapw/OmniRoute/pull/8921)) — thanks @Llliao1113

changelog.d/fixes/8960-fix.plan.md

@@ -1 +0,0 @@
-- fix(opencode): propagate vision capability from live catalog into opencode.json (#8960)

changelog.d/fixes/8965-fix.plan.md

@@ -1 +0,0 @@
-- fix(providers): switch Antigravity quota RPCs to iterate ANTIGRAVITY_RUNTIME_BASE_URLS (#8965)

changelog.d/fixes/8994-vertex-partner-claude.md

@@ -1 +0,0 @@
-- fix(vertex): route Claude models to native rawPredict endpoint and respect custom targetFormat overrides (#8994)

changelog.d/fixes/8995-fix.plan.md

@@ -1 +0,0 @@
-- fix(proxies): resolveProxyForConnection now returns the proxy name, so the dashboard badge shows the name instead of the hostname (#8995)

changelog.d/fixes/9029-cursor-narration.md

@@ -1 +0,0 @@
-- fix(cursor): preserve tool context across multi-turn conversations when client lacks conversation_id (#9029)

changelog.d/fixes/9030-antigravity-system-429s.md

@@ -1 +0,0 @@
-- fix(sse): move Antigravity client system content to first user message to avoid upstream 429 RESOURCE_EXHAUSTED on oversized systemInstruction (#9030)

changelog.d/fixes/9045-fix.plan.md

@@ -1 +0,0 @@
-- fix(db): stream DB backup export instead of buffering entire file into memory (#9045)

changelog.d/fixes/9046-fix.md

@@ -1 +0,0 @@
-- fix(ui): normalize Free Pool API response payload to read from data.proxies (#9046)

changelog.d/fixes/9054-fix.plan.md

@@ -1 +0,0 @@
-- fix(api/analytics): stop charging :free models at arbitrary fallback price (#9054)

changelog.d/fixes/9102-fix.plan.md

@@ -1 +0,0 @@
-- fix(providers): modal.com validation returns clear error when Base URL is missing, instead of leaking "Invalid outbound URL" (#9102)

changelog.d/fixes/9195-fix.plan.md

@@ -1,2 +0,0 @@
-- fix(catalog): repair dead guard and synced-first ordering for custom model Vision capable override (#9195)
-- fix(routing): consult customModels supportsVision flag in Combo vision filter (#9195)

changelog.d/fixes/9201-web-search-proxy-bind.plan.md

@@ -1 +0,0 @@
-- fix(web-search): bind each search provider attempt to its connection proxy (#9201)

changelog.d/fixes/9204-fix.plan.md

@@ -1 +0,0 @@
-- fix(auth): make antigravity and agy equivalent in credential selection (#9204)

changelog.d/fixes/9237-fix.plan.md

@@ -1 +0,0 @@
-- fix(lmarena): emit Uint8Array SSE chunks instead of strings to satisfy shared pipeline contract (#9237)

changelog.d/fixes/9277-fix.plan.md

@@ -1 +0,0 @@
-- fix(qoder): include actionable CLI_QODER_BIN hint in connection test when qodercli is not found (#9277)

changelog.d/fixes/9279-fix.plan.md

@@ -1 +0,0 @@
-- **fix(providers):** the web search fallback detector in `webSearchFallback.ts` used an exact `Set` (`web_search`, `web_search_preview`) that missed Anthropic's date-suffixed server-tool variant `web_search_20250305` (sent by Claude Code 2.1.220+). Changed to prefix regex `/^web_search/`, matching the two other detectors in the codebase, so the fallback intercepts versioned web search tools for OpenAI-compatible upstreams ([#9279](https://github.com/diegosouzapw/OmniRoute/pull/9279))

changelog.d/fixes/9289-fix.plan.md

@@ -1 +0,0 @@
-- fix(credential-health): scheduler never retries failed connections due to static interval comparison (#9289)

changelog.d/fixes/9293-fix.plan.md

@@ -1 +0,0 @@
-- fix(api): specialty model catalog ignores hidden OpenRouter model flags (#9293)

changelog.d/fixes/9300-fix.plan.md

@@ -1 +0,0 @@
-- fix(catalog): cache getModelsDevPricing() to prevent OOM at startup (#9300)

changelog.d/fixes/9304-fix.plan.md

@@ -1 +0,0 @@
-- fix(providers): bump qwen-web SPA version header from 0.2.66 to 0.2.81 (#9304)

changelog.d/fixes/9315-fix.plan.md

@@ -1 +0,0 @@
-- fix(backend): use accumulated responseBody for provider payload in dashboard log viewer to avoid stale data from truncated SSE events (#9315)

changelog.d/fixes/9319-fix.plan.md

@@ -1 +0,0 @@
-- fix(qoder): surface qodercli stderr in error message instead of generic 502 (#9319)

changelog.d/fixes/9431-codex-gpt56-context.md

@@ -1 +0,0 @@
-- **fix(providers):** Codex GPT-5.6 model metadata reports the 1M context window and 922K input limit ([#9431](https://github.com/diegosouzapw/OmniRoute/issues/9431)).

changelog.d/fixes/9494-chat-history-cap-opt-in.md

@@ -1 +0,0 @@
-- fix(api): make the 800-message chat history cap opt-in so long conversations reach compression instead of a terminal 413 (#9494)

changelog.d/fixes/9580-standalone-ws-multipart-uploads.md

@@ -1 +0,0 @@
-- **fix(standalone):** multipart uploads (`POST /v1/audio/transcriptions`) no longer hang — the WebDAV wrapper hands non-WebDAV requests to Next synchronously instead of losing the start of a streaming body ([#9580](https://github.com/diegosouzapw/OmniRoute/pull/9580))

changelog.d/fixes/9630-combo-false-503.md

@@ -1 +0,0 @@
-- fix(combo): distinguish pre-dispatch skips from genuine failures to prevent false 503 ALL_ACCOUNTS_INACTIVE (#9630)

changelog.d/fixes/9675-anonymous-fallback-disable-toggle.md

@@ -1 +0,0 @@
-- **fix(providers):** new per-provider `noAuthFallbackDisabledProviders` setting lets operators disable the synthetic anonymous (no-auth) credential fallback for API-key providers whose static definition declares `anonymousFallback: true` (e.g. `opencode-go`, `opencode-zen`) — upstream endpoints now reject anonymous requests with `401 Missing API key`, so the fallback added latency and caused UI health/reconnect churn. Real keyed connections keep working and recover automatically once quota state clears; true no-auth providers (`opencode`, `mimocode`, …) are unaffected, with `blockedProviders` remaining their disable mechanism. Default behavior is unchanged ([#9675](https://github.com/diegosouzapw/OmniRoute/pull/9675))

config/quality/eslint-suppressions.json

@@ -1356,11 +1356,24 @@
       "count": 1
     }
   },
+  "src/sse/handlers/chat.ts": {
+    "no-restricted-imports": {
+      "count": 1
+    }
+  },
   "src/sse/handlers/chatHelpers.ts": {
     "no-restricted-imports": {
       "count": 1
     }
   },
+  "src/sse/services/auth.ts": {
+    "no-restricted-imports": {
+      "count": 1
+    },
+    "no-restricted-syntax": {
+      "count": 1
+    }
+  },
   "src/sse/services/model.ts": {
     "no-restricted-imports": {
       "count": 2

config/quality/file-size-baseline.json

@@ -161,130 +161,185 @@
   "_rebaseline_2026_07_18_6979_codex_test": "PR #6979 own growth: executor-codex.test.ts 1340->1347 (+7 = generalized ensureThinkingBudget assertion added to the existing codex thinking-budget cases). antigravity-test bump 942->977 REVERTED here: #7408's test split dropped that file to 888, so this PR's +35 fits under the original 942 frozen cap.",
   "cap": 1000,
   "frozen": {
+    "_rebaseline_2026_07_02_5816_qoder": "PR #5816 (@AgentKiller45, qoder PAT via qodercli): qoderCli.ts 666->989, new-above-cap frozen (owner-approved baseline freeze). The growth is the legitimate PAT job-token exchange + quota parsing CLI transport (the pure-JS Cosy path 500'd on every PAT request); extracting the spawn/parse helpers now would just add indirection to a contributor PR mid-merge. Test frozen also raised for this PR's coverage growth: providers-page-utils.test.ts 1052->1092. Additionally clears an inherited base-red from the already-merged #5933 (codex json_schema->text.format): translator-openai-responses-req.test.ts 1097->1172 (+75 regression tests, no offending branch left). All remain frozen (cannot grow further); release captain's rebaseline-at-release supersedes.",
+    "open-sse/services/qoderCli.ts": 989,
+    "_rebaseline_pr1043_minimax_tts": "Upstream port decolua/9router#1043 (toanalien) own growth: audioSpeech.ts 965->1061 (+96). Adds MiniMax T2A v2 TTS dispatch (handleMinimaxSpeech + hexToBytes helper) — provider entry was already in audioRegistry (format: minimax-tts) but no handler existed, falling through to the OpenAI-compatible default that fails (T2A has custom shape + hex-encoded audio + base_resp envelope). New branch sits next to the other inline provider branches (xiaomi-mimo, coqui, tortoise, aws-polly) — extracting would just create indirection. Covered by tests/unit/minimax-tts-1043.test.ts (3 tests, GREEN: success, base_resp error, invalid-hex).",
+    "_rebaseline_2026_07_23_8266_alibaba_media": "#8266 (@backryun) own growth: imageRegistry.ts 821->979 (+158) — Alibaba-family media models (Qwen image/video, Bailian, Wan) added to the image/video registry. Registry model data, not extractable logic; frozen at new size.",
+    "open-sse/config/imageRegistry.ts": 979,
+    "open-sse/config/providerRegistry.ts": 4731,
+    "open-sse/executors/antigravity.ts": 1813,
+    "open-sse/executors/base.ts": 1540,
+    "open-sse/executors/chatgpt-web.ts": 3206,
+    "open-sse/executors/claude-web.ts": 1057,
+    "open-sse/executors/codex.ts": 1541,
+    "open-sse/executors/cursor.ts": 1577,
+    "open-sse/executors/deepseek-web.ts": 1148,
     "_rebaseline_2026_06_22_4644_deepseek_web_tools": "PR #4644 (BugsBag/robust deepseek-web tool-call parsing): open-sse/executors/deepseek-web.ts 1117->1125 (+8). The new agentic tool-call path emits surrounding text + reasoning before tool_calls and swaps to the dedicated deepseekWebTools.ts parser; the +8 lines are cohesive wiring at the existing transformSSE chokepoint (the parser itself lives in the new deepseekWebTools.ts file, already under cap). The PR's own fast-gate (PR->release) does not run check:file-size, so this surfaced only at release reconcile. Covered by tests/unit/deepseek-web-tools-variants.test.ts + deepseek-web-tools-execute.test.ts.",
     "_rebaseline_2026_06_23_4712_deepseek_web_tool_results": "PR for #4712 (deepseek-web drops role:tool): open-sse/executors/deepseek-web.ts 1125->1148 (+23). messagesToPrompt() now folds role:\"tool\" results into the single-prompt transcript (recovering the tool name from the preceding assistant tool_calls by tool_call_id) instead of silently dropping them; the lines are cohesive wiring inside the existing function. Covered by tests/unit/deepseek-web-tool-result-prompt-4712.test.ts.",
+    "_rebaseline_2026_06_28_5237_impersonation_ua_refresh": "PR #5237 (refresh impersonation UAs): grok-web.ts 1871->1873 (+2), muse-spark-web.ts 1284->1302 (+18), perplexity-web.ts 1013->1032 (+19). Net semantic change in each file is a single User-Agent constant (Chrome 147->149 for grok/muse; perplexity kept at Firefox 148 to stay matched with the firefox_148 TLS profile — the contributor's 152 bump was reverted to avoid a UA-vs-JA3 mismatch, #2459). The growth is Prettier reflow that lint-staged unavoidably applies to these grandfathered long-line files the moment they are touched; not extractable. src/sse/services/auth.ts 2336->2401 in the same reconcile is #5222's antigravity-LRU-retry growth that merged via --admin without a baseline bump.",
+    "open-sse/executors/duckduckgo-web.ts": 925,
+    "open-sse/executors/grok-web.ts": 1873,
+    "open-sse/executors/hyperagent.ts": 937,
+    "open-sse/executors/muse-spark-web.ts": 1396,
+    "open-sse/executors/perplexity-web.ts": 1032,
+    "open-sse/handlers/audioSpeech.ts": 1061,
+    "open-sse/handlers/chatCore.ts": 5125,
+    "open-sse/handlers/imageGeneration.ts": 3777,
+    "open-sse/handlers/responseSanitizer.ts": 1139,
+    "open-sse/handlers/search.ts": 1546,
+    "open-sse/handlers/sseParser.ts": 830,
+    "open-sse/handlers/videoGeneration.ts": 1275,
+    "_rebaseline_2026_07_22_8010_codex_responses_engine": "PR #8010 (@JxnLexn) own growth: open-sse/mcp-server/schemas/tools.ts 1497->1505 (+8 = threading the new \"codex-responses\" literal into the compressionConfigureInput strategy/autoTriggerMode Zod enums and setCompressionEngineInput engine enum, mirroring the existing rtk/omniglyph enum entries; no new tool). open-sse/services/compression/strategySelector.ts 1043->1054 (+11 = one new `if (mode === \"codex-responses\")` dispatch branch in runCompression that delegates 100% to the new codexResponsesEngine.apply, mirroring the existing rtk single-mode dispatch, plus threading config.codexResponsesConfig.preserveToolNames into the shared adaptBodyForCompression call at the 3 existing call sites). src/lib/db/compression.ts (untracked, new-file cap 800) 794->845 (+51 = normalizeCodexResponsesConfig, mirroring the existing normalizeRtkConfig normalizer, plus registering \"codex-responses\" in the COMPRESSION_MODES/STACKED_PIPELINE_ENGINE_IDS/SINGLE_MODE_ENGINE sets and the getCompressionSettings load/save switch) — added to the baseline at its current size. All three are cohesive dispatch/normalizer wiring at existing chokepoints (mirroring the prior compression-mode rebaselines #6534/#6556), not extractable without hiding the mode-dispatch boundary. Covered by tests/unit/compression/codex-responses.test.ts (6) + omniglyph-registries.test.ts/types.test.ts (22, updated for the new mode).",
+    "_rebaseline_2026_07_22_8034_compression_exclusions_persistence": "#8034 (compression exclusions) own growth: src/lib/db/compression.ts 845->850 (+5 = threading the new compressionExclusions field through the existing getCompressionSettings/saveCompressionSettings load/save switch over the shared key_value compression namespace — no new table, no raw SQL). Mirrors the prior compression-field rebaselines (#8010 codex-responses normalizer at the same chokepoint); the load/save switch is a single dispatch boundary, not extractable without hiding it. Covered by the PR's 8 node:test + 3 vitest cases.",
+    "src/lib/db/compression.ts": 866,
+    "open-sse/mcp-server/schemas/tools.ts": 1505,
+    "open-sse/mcp-server/server.ts": 1555,
+    "open-sse/mcp-server/tools/advancedTools.ts": 1120,
+    "_rebaseline_2026_06_27_5193_antigravity_basered": "Base-red (pre-existing release drift, fast-gate PR->release skips check:file-size): accountFallback.ts 1773->1777 and src/app/api/providers/[id]/test/route.ts 924->940 were already over their frozen caps on release/v3.8.39 independent of any antigravity change. Owner chose to rebaseline (keep the documented issue-reference comments #1846/#1449/#347 etc.) rather than accept the contributor comment-stripping in #5200/#5198. Reverted #5200 to restore the comments; bumped these two frozen caps to the actual base sizes. No logic change.",
+    "_rebaseline_2026_07_22_8213_gemini_tpm_quota_cooldown_wait": "PR #8213 (hartmark, fix/gemini-tpm-quota-cooldown-wait) own growth: open-sse/services/accountFallback.ts 1857->1932 on the merged tip (release 1892 incl #8050 +35, plus this PR own growth +40); measured against the PR own merge-base was 1857->1898 (+41 — the release tip separately carries an unrelated +34 from #8050's antigravity 404 model-not-found lockout scoping, which this PR's branch does not include and this entry does not cover). Own growth is the Gemini TPM-ceiling classification + cooldown-wait wiring feeding into the combo cooldown-wait state machine (rate-limit wedge recovery) introduced by this PR's commit series. Irreducible additions at the existing account-fallback/model-lockout chokepoint. Covered by the PR's own gemini-rate-limit-tracker and TPM-ceiling benchmark test additions.",
+    "_rebaseline_2026_07_23_8252_combo_400_advance": "#8252 (@RaviTharuma) own growth: accountFallback.ts 1932->1940 (+8) + combo.ts 3604->3630 (+26) — advance combo on model-scoped 400s wrapped as invalid/Bad-Request. Irreducible wiring at existing account-fallback + combo dispatch chokepoints. Covered by combo-model-scoped-400-advance.test.ts.",
+    "open-sse/services/accountFallback.ts": 1940,
+    "open-sse/services/adobeFireflyClient.ts": 1958,
+    "open-sse/services/batchProcessor.ts": 915,
+    "open-sse/services/browserBackedChat.ts": 850,
+    "open-sse/services/claudeCodeCompatible.ts": 1202,
+    "_rebaseline_pr4592_exclude_exhausted_auto": "Reconcile #4592 already-merged growth: combo.ts 2991->3036 (+45, terminal-status quota-cutoff exclusion in buildAutoCandidates + opt-in gate). Fast-gate PR->release does not run check:file-size.",
     "_rebaseline_2026_06_24_headroom_strategy": "Headroom-aware connection selection (dario technique): combo.ts 3168->3180 (+12 = a new `else if (strategy === \"headroom\")` dispatch branch in handleComboChat that delegates to orderTargetsByHeadroom + its log line, plus the import). The actual logic lives OUT of the god-file: the pure ranker rankByHeadroom/computeHeadroom is the new leaf open-sse/services/combo/headroomRanking.ts (91 LOC, <cap) and the async orderer orderTargetsByHeadroom is appended to the existing open-sse/services/combo/quotaStrategies.ts (<cap) next to its sibling reset-aware/reset-window orderers (reuses their connection-expansion machinery). headroom = 1 - max(util_5h, util_7d) from getSaturation (src/lib/quota/saturationSignals.ts), prefers the connection with the most free capacity. Only the dispatch wiring is irreducible at the existing combo strategy chokepoint (mirrors the reset-aware/reset-window/context-optimized branches); not extractable without hiding the call site. fill-first stays default; all existing strategies untouched. Covered by tests/unit/combo-headroom-ranking.test.ts (pure helper) + tests/unit/combo-headroom-strategy.test.ts (orderer, saturation injected). Structural shrink of combo.ts tracked in #3501.",
     "_rebaseline_2026_06_24_quota_share_strategy": "Dedicated quota-share strategy (Phase 3 #9): combo.ts 3180->3190 (+10 = one new `else if (strategy === \"quota-share\")` dispatch branch in handleComboChat that delegates 100% to selectQuotaShareTarget + its log line, plus the import). All the new logic lives OUT of the god-file in two new leaves under open-sse/services/combo/: quotaShareInflight.ts (in-flight counter with TTL/lease, ~150 LOC <cap) and quotaShareStrategy.ts (per-model bucket gating via isBucketSaturated + DRR proportional to weight + P2C over in-flight, ~240 LOC <cap). Only the dispatch wiring is irreducible at the existing combo strategy chokepoint (mirrors the headroom/reset-aware/reset-window/context-optimized branches); not extractable without hiding the call site. ZERO existing strategy cases were modified — only this branch was added, and the qtSd/ combos switched from fill-first to quota-share in src/lib/quota/quotaCombos.ts. Covered by tests/unit/quota-share-strategy.test.ts (gating, DRR fairness, P2C in-flight, fail-open, activation). Structural shrink of combo.ts tracked in #3501.",
     "_rebaseline_2026_06_24_task_aware_routing": "Task-aware routing strategy (port PR #2045, OmniRoute #4945): combo.ts 3190->3225 (+35) = one new `else if (strategy === \"task-aware\")` dispatch branch delegating 100% to selectTaskAwareTarget + its imports/log lines. All scoring/classification logic lives OUT of the god-file in the new leaf open-sse/services/taskAwareRouting.ts (553 LOC <cap). Only the dispatch wiring is irreducible at the existing combo strategy chokepoint (mirrors quota-share/headroom/reset-aware branches). ZERO existing strategy cases modified. Covered by tests/unit/combo-task-aware.test.ts (35 tests). Structural shrink of combo.ts tracked in #3501.",
+    "_rebaseline_2026_07_22_8213_combo_cooldown_wait_recording": "PR #8213 (hartmark, fix/gemini-tpm-quota-cooldown-wait) own growth: open-sse/services/combo.ts 3548->3604 (+56, entirely this PR's diff — no other commit touched this file between the PR's merge-base and the release tip). Fixes combo cooldown-wait state recording so a bogus 503 is no longer crystallized when the cooldown-wait vars reset every setTry, adds an OpenAI-format SSE error frame path for combo-exhausted rejections (capturing request body + attempted models), and gives an abandoned per-target dispatch its own timeout instead of leaking a permanent 'pending' dashboard entry. Irreducible additions at the existing handleComboChat dispatch/retry chokepoint (mirrors the prior quota-share/headroom/task-aware strategy-branch precedents already frozen in this file). Covered by the PR's own combo-config + Gemini TPM-ceiling benchmark test additions.",
+    "open-sse/services/combo.ts": 3630,
     "_rebaseline_2026_06_26_fidelity_gate_extraction": "Milestone-B fidelity-gate wiring residual: bodyToText+gateAdvance extracted to fidelityGateStep.ts (889->854, -35), but the StackOptions.fidelityGate field, the `const fidelityGate` reads at the two stacked-loop dispatch chokepoints, and the import of FidelityGateConfig are irreducible wiring that cannot leave strategySelector without an architectural refactor of the pre-existing stacked pipeline. Net: 889->854 (+6 vs the pre-Milestone-B frozen 848). Covered by tests/unit/compression/*.test.ts (940 pass).",
-    "_rebaseline_2026_06_27_5193_5203_antigravity_oauthmodal": "Antigravity remote-login own growth: OAuthModal.tsx 960->969 (gate units). #5193 (+~4: remote paste instruction shown for all remote incl. Google + its rationale comment) and #5203 (+~5: handleManualSubmit credential-blob branch + button guard; submit logic extracted to oauthBlobSubmit.ts to minimize). Frozen set to the SUM so either merge order passes. Cohesive at the existing manual-submit chokepoint.",
-    "_rebaseline_2026_06_27_5193_antigravity_basered": "Base-red (pre-existing release drift, fast-gate PR->release skips check:file-size): accountFallback.ts 1773->1777 and src/app/api/providers/[id]/test/route.ts 924->940 were already over their frozen caps on release/v3.8.39 independent of any antigravity change. Owner chose to rebaseline (keep the documented issue-reference comments #1846/#1449/#347 etc.) rather than accept the contributor comment-stripping in #5200/#5198. Reverted #5200 to restore the comments; bumped these two frozen caps to the actual base sizes. No logic change.",
-    "_rebaseline_2026_06_28_5237_impersonation_ua_refresh": "PR #5237 (refresh impersonation UAs): grok-web.ts 1871->1873 (+2), muse-spark-web.ts 1284->1302 (+18), perplexity-web.ts 1013->1032 (+19). Net semantic change in each file is a single User-Agent constant (Chrome 147->149 for grok/muse; perplexity kept at Firefox 148 to stay matched with the firefox_148 TLS profile — the contributor's 152 bump was reverted to avoid a UA-vs-JA3 mismatch, #2459). The growth is Prettier reflow that lint-staged unavoidably applies to these grandfathered long-line files the moment they are touched; not extractable. src/sse/services/auth.ts 2336->2401 in the same reconcile is #5222's antigravity-LRU-retry growth that merged via --admin without a baseline bump.",
     "_rebaseline_2026_06_28_5243_risk_gate_prepass": "PR #5243 (compression risk-gate pre-pass) own growth: open-sse/services/compression/strategySelector.ts 854->899 (+45). The three exported entry points (applyCompression/applyStackedCompression/applyStackedCompressionAsync) become thin wrappers over pure-extracted private bodies (runCompression/runStackedCompression/runStackedCompressionAsync) so the risk-gate mask->run->restore wrapper sits strictly OUTSIDE the per-step loop — a single universal integration point. The wrapper logic itself (resolveRiskGate/withRiskGate) lives in the new riskGate/strategyWrap.ts (<cap); the residual growth is the duplicated thin-wrapper signatures + the extracted bodies' dispatch boundary, guarded by a byte-identical parity test (riskGateIntegration). Default off (DEFAULT_COMPRESSION_CONFIG unchanged). Not extractable without hiding the dispatch boundary, mirroring prior compression rebaselines. Structural shrink tracked in #3501.",
-    "_rebaseline_2026_06_28_5275_correlation_id_extract": "Extraction of the safe CorrelationId subset of #5275 (hartmark) — request correlation id stored in call_logs (migration 109) and returned via the X-Correlation-Id response header, WITHOUT the combo/resilience or build/lazy-loading changes (those stay in #5275). Own growth: callLogs.ts 975->985 (correlation_id column on CallLogSummaryRow + read/map), usageHistory.ts 983->988 (correlationId metadata normalize), chat.ts 1575->1632 (withCorrelationId response wiring + combo-failure log carrying correlationId), chatHelpers.ts new 811 (withCorrelationId helper + reqId threading; was 791<cap pre-feature). Cohesive request/logging chokepoint wiring; structural shrink of chat.ts tracked in #3501.",
-    "_rebaseline_2026_06_29_4038_cas_guard": "PR (#4038) own growth: tokenRefresh.ts 2103->2181 (+78 = the compare-and-swap guard on the refresh persist — runWithCasGuard/getActiveCasGuard AsyncLocalStorage pair mirroring runWithOnPersist, casGuardShouldSkipPersist that rereads the row right before persisting and skips the write when a concurrent writer already rotated the refresh_token past the one presented, plus getCasGuardStats counters). Fixes the sibling-rotation-revert → token-family-revocation storm. Gated behind an active guard (opt-in; no guard => byte-identical). Wiring lives at the two persist chokepoints inside getAccessToken; the comparison reuses wasRefreshTokenRotated from refreshSerializer. Not extractable without splitting the refresh hot path.",
     "_rebaseline_2026_06_29_5286_memoization": "PR #5286 own growth: strategySelector.ts 899->960 (+61 = the opt-in result-memoization branches in applyCompression/applyCompressionAsync — principal+determinism gate, makeMemoKey lookup/store with model+supportsVision folded into the key, recompute-with-memo-off). Default off (memoizeCompressionResults), so zero behavior change. The memo helpers live in the leaf resultMemo.ts (<cap); the chokepoint wiring here is not extractable. Structural shrink of this hot-path file tracked in #3501.",
-    "_rebaseline_2026_07_01_v3843_release_5609": "Rebaseline v3.8.43 (PR #5609 release reconciliation). DRIFT dos 109 commits do ciclo: 8 god-files existentes cresceram (ApiManagerPageClient 2983->3017, combos/page 4594->4608, AddApiKeyModal 868->869, providerPageHelpers 974->996, chat.ts 1635->1647, auth.ts 2401->2403, batchProcessor 828->915, combo.ts 3368->3387) + 2 novos acima do cap (huggingchat.ts 813, tests web-cookie-providers-new 827) + 4 test files cresceram. Modularizacao deferida (blast-radius mid-release); congelado no estado atual p/ o proximo ciclo ratchetar daqui.",
-    "_rebaseline_2026_07_02_5816_qoder": "PR #5816 (@AgentKiller45, qoder PAT via qodercli): qoderCli.ts 666->989, new-above-cap frozen (owner-approved baseline freeze). The growth is the legitimate PAT job-token exchange + quota parsing CLI transport (the pure-JS Cosy path 500'd on every PAT request); extracting the spawn/parse helpers now would just add indirection to a contributor PR mid-merge. Test frozen also raised for this PR's coverage growth: providers-page-utils.test.ts 1052->1092. Additionally clears an inherited base-red from the already-merged #5933 (codex json_schema->text.format): translator-openai-responses-req.test.ts 1097->1172 (+75 regression tests, no offending branch left). All remain frozen (cannot grow further); release captain's rebaseline-at-release supersedes.",
+    "open-sse/services/compression/strategySelector.ts": 1060,
+    "_rebaseline_2026_07_10_gcf_v3_2_decode": "PR #6838 own growth: new vendored file open-sse/services/compression/engines/headroom/gcf/decode_generic.ts frozen at 880 (> 800 cap). It is the vendored GCF generic-profile decoder (spec v3.2 nested flattening plus the prototype-pollution / hasOwnProperty hardening added in this PR's Gemini review). Kept as one file faithful to upstream gcf-typescript so re-vendoring stays a clean copy rather than a re-split each cycle (sibling generic.ts/scalar.ts stay < cap; extraction would also fragment the file's frozen eslint no-explicit-any suppressions). Round-trip + prototype-pollution regression coverage in tests/unit/compression/headroom-smartcrusher.test.ts. Frozen: only shrinks from here.",
+    "open-sse/services/compression/engines/headroom/gcf/decode_generic.ts": 880,
+    "open-sse/services/rateLimitManager.ts": 1035,
+    "_rebaseline_2026_06_29_4038_cas_guard": "PR (#4038) own growth: tokenRefresh.ts 2103->2181 (+78 = the compare-and-swap guard on the refresh persist — runWithCasGuard/getActiveCasGuard AsyncLocalStorage pair mirroring runWithOnPersist, casGuardShouldSkipPersist that rereads the row right before persisting and skips the write when a concurrent writer already rotated the refresh_token past the one presented, plus getCasGuardStats counters). Fixes the sibling-rotation-revert → token-family-revocation storm. Gated behind an active guard (opt-in; no guard => byte-identical). Wiring lives at the two persist chokepoints inside getAccessToken; the comparison reuses wasRefreshTokenRotated from refreshSerializer. Not extractable without splitting the refresh hot path.",
     "_rebaseline_2026_07_09_6126_clinepass_dual_auth": "PR #6126 (@hajilok, dual-auth ClinePass) own growth: tokenRefresh.ts 2181->2182 (+1 = a single `case \"clinepass\":` fallthrough label added to the existing `case \"cline\":` in _getAccessTokenInternal's provider switch, so clinepass token refresh dispatches to the already-shared refreshClineToken() instead of silently falling through to the generic OAuth refresh). Irreducible 1-line switch-case wiring at the existing chokepoint; the header-building logic for the same feature was extracted to a new leaf src/shared/utils/clineAuth.ts::buildClinepassHeaders() (well under cap) to avoid growing open-sse/executors/default.ts. Covered by tests/unit/clinepass-provider.test.ts.",
     "_rebaseline_2026_07_09_6363_kiro_external_idp": "PR #6363 (@artickc, Kiro external IdP) own growth: tokenRefresh.ts 2182->2249 (+67 = the external_idp refresh branch inside refreshKiroToken — standard public-client OAuth2 refresh_token grant against the org IdP tokenEndpoint via buildExternalIdpRefreshParams/isExternalIdpAuthMethod from the new leaf open-sse/services/kiroExternalIdp.ts, with invalid_grant/invalid_client -> unrecoverable_refresh_error mapping). Cohesive addition at the existing refreshKiroToken chokepoint. Covered by tests/unit/kiro-external-idp.test.ts.",
-    "_rebaseline_2026_07_09_6587_kiro_api_key_auth": "PR #6587 (@strangersp) own growth for Kiro long-lived API-key auth, merged onto v3.8.47 tip: openai-to-kiro.ts 890->912 (+22, auth-header selection for API-key-vs-OAuth-token connections), providerLimits.ts 998->1000 (+2, API-key auth-type branch), translator-openai-to-kiro.test.ts 1234->1257 (+23), providers-page-utils.test.ts 1109->1107 (net -2 after merging with parallel release drift; connectionMatchesProviderCard api_key coverage added), provider-validation-specialty.test.ts 2856->2980 (+124 net after merge with parallel release drift; this PR also removed the file's `@typescript-eslint/no-explicit-any` eslint-suppression entry by fixing all `any` usages, adding typed replacements). Cohesive additive feature growth, well tested; not extractable without splitting the existing chokepoints mid-merge.",
-    "_rebaseline_2026_07_09_6678_routing_strategy_9router": "#6678 (SeaXen) — 9router-parity Routing Strategy settings card + per-provider/combo sticky-round-robin override. Own growth: ProviderDetailPageClient.tsx 784->786 (single ProviderAccountRoutingCard mount + import), auth.ts 2448->2458 (providerStrategies override resolution: fallbackStrategy/stickyRoundRobinLimit per-provider cascade in getProviderCredentials). Both additive, zero unrelated refactor; new UI/logic lives in new files (ProviderAccountRoutingCard.tsx, RoutingStrategyCard.tsx, rrState.ts::resolveComboStickyRoundRobinLimit). chat.ts value below reflects the current release tip (grown by other concurrent PRs, e.g. #6640), not this PR own change.",
-    "_rebaseline_2026_07_10_6318_omp_letta": "PR #6318 (@hamsa0x7, omp+letta CLI integrations) own growth: cliTools.ts (+53 = 2 registry entries incl. omp docsUrl) and cliRuntime.ts (+18 = runtime-detection wiring for the 2 new tools). Cohesive registry/wiring growth at the existing chokepoints; scope reduced from the original 5 tools (pi/codewhale/jcode shipped separately).",
-    "_rebaseline_2026_07_10_gcf_v3_2_decode": "PR #6838 own growth: new vendored file open-sse/services/compression/engines/headroom/gcf/decode_generic.ts frozen at 880 (> 800 cap). It is the vendored GCF generic-profile decoder (spec v3.2 nested flattening plus the prototype-pollution / hasOwnProperty hardening added in this PR's Gemini review). Kept as one file faithful to upstream gcf-typescript so re-vendoring stays a clean copy rather than a re-split each cycle (sibling generic.ts/scalar.ts stay < cap; extraction would also fragment the file's frozen eslint no-explicit-any suppressions). Round-trip + prototype-pollution regression coverage in tests/unit/compression/headroom-smartcrusher.test.ts. Frozen: only shrinks from here.",
-    "_rebaseline_2026_07_12_v3847_mergeprs_tail": "v3.8.47 /merge-prs tail (owner-approved): src/lib/localDb.ts NEW>800 (799->805, +6 re-exports countFreeProxies + recordFreeProxySyncErrors/clearFreeProxySyncErrors/getFreeProxySyncErrors + FreeProxySyncErrors type for #6909 free-pool relay-repair; re-export-only per Hard Rule #2, not extractable).",
+    "open-sse/services/tokenRefresh.ts": 2249,
+    "open-sse/services/usage.ts": 3454,
+    "open-sse/translator/request/openai-to-gemini.ts": 906,
+    "open-sse/translator/request/openai-to-kiro.ts": 912,
+    "_rebaseline_2026_07_22_8211_gemini_malformed_tool_choice": "PR #8211 (hartmark, fix/gemini-malformed-function-call-tool-choice) own growth: open-sse/translator/response/gemini-to-openai.ts 771->821 (+50, entirely this PR's diff — no other commit touched this file between the PR's merge-base and the release tip). Adds MALFORMED_FUNCTION_CALL/UNEXPECTED_TOOL_CALL handling inside geminiToOpenAIResponse(): synthesizes a `malformed_tool_call` tool_calls entry so finish_reason normalizes to the standard \"tool_calls\" instead of an unrecognized raw enum value that OpenAI-compatible clients (e.g. OpenClaw) silently ignore, and always synthesizes (rather than skipping when a real tool call already exists) so a malformed attempt alongside a real one in the same turn is not silently discarded. Irreducible cohesive addition at the existing candidate/finishReason translation chokepoint (mirrors the 9router#2462 raw-finish-reason precedent immediately below it in the same function). Covered by the PR's own tests/unit test additions for both the malformed-only and malformed-plus-real-call cases.",
+    "open-sse/translator/response/gemini-to-openai.ts": 821,
+    "_rebaseline_2026_07_22_7936_namespace_roundtrip": "#7936 (@RCrushMe, Responses-Chat namespace round-trip identity seam) own growth: open-sse/translator/response/openai-responses.ts 1092->1125 (+33) and open-sse/utils/stream.ts 2814->2869 (+55) — threading the namespace-identity seam through the Responses↔Chat translation + stream paths so tool-call namespaces survive the round-trip. Cohesive translation/stream wiring at existing chokepoints, frozen at new size.",
+    "_rebaseline_2026_07_22_8210_openrouter_midstream_error": "PR #8210 (hartmark, fix/openrouter-midstream-error-surfacing) own growth: open-sse/translator/response/openai-responses.ts 1137->1163 (+26) measured on the merged tip (release 1137 + this PR own growth). Adds a single new branch inside openaiToOpenAIResponsesResponse() that detects an OpenRouter-style mid-stream aggregator error (HTTP 200 SSE chunk with empty choices + a top-level error object) and surfaces it as state.upstreamError instead of silently falling through to the no-op/awaitingTrailingUsage path, which previously masked the failure as a false empty-success completion and skipped combo fallback. Irreducible call-site addition at the existing chunk-dispatch chokepoint (mirrors the Gemini-to-OpenAI translator's #4177 precedent for the same class of upstream error surfacing). Note: this baseline entry does NOT cover the separate pre-existing +11 drift already on the release tip from #8081/#8162 (1125->1136, unrelated reasoning-placeholder-stripping fix merged after this PR branched) — that drift belongs to the maintainer's rebaseline, not this PR.",
+    "open-sse/translator/response/openai-responses.ts": 1163,
+    "open-sse/utils/cursorAgentProtobuf.ts": 1521,
+    "_rebaseline_2026_07_23_8143_empty_catch_logging": "#8143 (@chirag127) own growth: open-sse/utils/stream.ts 2869->2887 (+18) — replacing empty catch blocks in the SSE stream subsystem with console.debug logging (Rule #6 silent-swallow fix, issues #8138-#8142). Cohesive logging additions at the existing catch chokepoints, not extractable; frozen at new size. Covered by tests/unit/stream-handler-catch-logging-8143.test.ts.",
+    "open-sse/utils/stream.ts": 2887,
+    "src/app/(dashboard)/dashboard/HomePageClient.tsx": 1385,
+    "src/app/(dashboard)/dashboard/analytics/ComboHealthTab.tsx": 1031,
+    "src/app/(dashboard)/dashboard/api-manager/ApiManagerPageClient.tsx": 3120,
+    "src/app/(dashboard)/dashboard/cache/media/MediaPageClient.tsx": 1105,
+    "src/app/(dashboard)/dashboard/cache/page.tsx": 845,
+    "src/app/(dashboard)/dashboard/cli-code/components/CodexToolCard.tsx": 900,
+    "src/app/(dashboard)/dashboard/cloud-agents/page.tsx": 931,
     "_rebaseline_2026_07_15_7070_combos_memo": "PR #7070 (perf/p1-memo) own growth: src/app/(dashboard)/dashboard/combos/page.tsx 4655->4656 (+1 = React.memo wrapping of ComboCard). Covered by tests/unit/ui/combos-page-smoke.test.tsx.",
+    "src/app/(dashboard)/dashboard/combos/page.tsx": 4656,
+    "src/app/(dashboard)/dashboard/costs/CostOverviewTab.tsx": 1495,
+    "src/app/(dashboard)/dashboard/costs/quota-share/components/PoolWizard.tsx": 1022,
+    "src/app/(dashboard)/dashboard/endpoint/EndpointPageClient.tsx": 2615,
+    "_rebaseline_2026_07_22_8213_health_unblock_model_cooldowns": "PR #8213 (hartmark, fix/gemini-tpm-quota-cooldown-wait) own growth: src/app/(dashboard)/dashboard/health/page.tsx 1094->1165 (+71, entirely this PR's diff — no other commit touched this file between the PR's merge-base and the release tip). Adds handleUnblockAll/handleUnblockOne dashboard actions (DELETE /api/resilience/model-cooldowns) so an operator can manually clear a Gemini TPM-wedge model lockout surfaced by this PR's cooldown-wait fixes, instead of waiting out the ceiling. Irreducible UI wiring at the existing health-page action chokepoint. Covered by the PR's own dashboard/resilience test additions.",
+    "src/app/(dashboard)/dashboard/health/page.tsx": 1165,
+    "src/app/(dashboard)/dashboard/playground/components/tabs/ApiTab.tsx": 847,
+    "src/app/(dashboard)/dashboard/providers/[id]/ProviderDetailPageClient.tsx": 804,
+    "src/app/(dashboard)/dashboard/providers/[id]/components/ConnectionRow.tsx": 958,
+    "src/app/(dashboard)/dashboard/providers/[id]/components/modals/AddApiKeyModal.tsx": 967,
+    "src/app/(dashboard)/dashboard/providers/[id]/components/modals/EditConnectionModal.tsx": 1288,
+    "src/app/(dashboard)/dashboard/providers/[id]/hooks/useProviderConnections.ts": 986,
+    "src/app/(dashboard)/dashboard/providers/[id]/hooks/useProviderModels.ts": 155,
+    "src/app/(dashboard)/dashboard/providers/[id]/hooks/useProviderSettings.ts": 264,
+    "src/app/(dashboard)/dashboard/providers/[id]/providerPageHelpers.ts": 1054,
+    "src/app/(dashboard)/dashboard/providers/components/onboarding/ProviderOnboardingWizard.tsx": 948,
+    "src/app/(dashboard)/dashboard/providers/page.tsx": 1927,
+    "src/app/(dashboard)/dashboard/runtime/RuntimePageClient.tsx": 1201,
+    "src/app/(dashboard)/dashboard/settings/components/AppearanceTab.tsx": 819,
+    "src/app/(dashboard)/dashboard/settings/components/ComboDefaultsTab.tsx": 903,
+    "src/app/(dashboard)/dashboard/settings/components/CompressionSettingsTab.tsx": 974,
+    "src/app/(dashboard)/dashboard/settings/components/MemorySkillsTab.tsx": 898,
+    "src/app/(dashboard)/dashboard/settings/components/PricingTab.tsx": 1019,
+    "src/app/(dashboard)/dashboard/settings/components/ProxyRegistryManager.tsx": 1464,
+    "src/app/(dashboard)/dashboard/settings/components/ResilienceTab.tsx": 1183,
+    "src/app/(dashboard)/dashboard/settings/components/RoutingTab.tsx": 1629,
+    "src/app/(dashboard)/dashboard/settings/components/SystemStorageTab.tsx": 1924,
+    "src/app/(dashboard)/dashboard/usage/components/BudgetTab.tsx": 1028,
+    "src/app/(dashboard)/dashboard/usage/components/EvalsTab.tsx": 2148,
+    "src/app/(dashboard)/dashboard/usage/components/ProviderLimits/index.tsx": 1127,
+    "src/app/api/oauth/[provider]/[action]/route.ts": 970,
+    "src/app/api/providers/[id]/models/route.ts": 2593,
+    "src/app/api/providers/[id]/test/route.ts": 940,
+    "src/app/api/usage/analytics/route.ts": 948,
+    "src/app/api/v1/models/catalog.ts": 1615,
+    "src/lib/cloudflaredTunnel.ts": 935,
+    "src/lib/db/apiKeys.ts": 1662,
+    "src/lib/db/core.ts": 1825,
+    "src/lib/db/migrationRunner.ts": 1125,
+    "src/lib/db/models.ts": 1259,
+    "src/lib/db/providers.ts": 1107,
+    "src/lib/db/proxies.ts": 1177,
+    "src/lib/db/settings.ts": 1155,
+    "src/lib/db/usageAnalytics.ts": 925,
+    "src/lib/evals/evalRunner.ts": 961,
+    "src/lib/memory/retrieval.ts": 1171,
+    "src/lib/modelsDevSync.ts": 934,
+    "src/lib/providers/validation.ts": 4523,
+    "src/lib/resilience/settings.ts": 841,
+    "src/lib/tailscaleTunnel.ts": 1202,
+    "src/lib/usage/callLogs.ts": 997,
+    "src/lib/usage/providerLimits.ts": 1006,
+    "src/lib/usage/usageHistory.ts": 988,
+    "_rebaseline_2026_06_27_5193_5203_antigravity_oauthmodal": "Antigravity remote-login own growth: OAuthModal.tsx 960->969 (gate units). #5193 (+~4: remote paste instruction shown for all remote incl. Google + its rationale comment) and #5203 (+~5: handleManualSubmit credential-blob branch + button guard; submit logic extracted to oauthBlobSubmit.ts to minimize). Frozen set to the SUM so either merge order passes. Cohesive at the existing manual-submit chokepoint.",
     "_rebaseline_2026_07_18_7399_xai_oauth_modal": "PR #7399 (xAI OAuth PKCE) own growth: OAuthModal.tsx 993->998 (+5 = provider entry + PKCE flow branch wiring at the existing provider-switch chokepoint; the provider logic itself lives in src/lib/oauth/providers/xai-oauth.ts, new leaf). Third irreducible wiring bump on this modal (969->989->993->998); structural shrink tracked in #3501.",
     "_rebaseline_2026_07_19_6636_codex_session_json": "#6636 own growth: OAuthModal.tsx 998->1030 (gate units, split(\"\\n\").length incl. trailing newline; +32 = session-JSON paste branch for handleManualSubmit plus a shared submitCodexAccessToken() helper extracted from the pre-existing bare-JWT branch, mirroring the #5203 oauthBlobSubmit.ts extraction precedent; the normalizer logic itself lives in the new src/lib/oauth/utils/codexSessionImport.ts leaf module, not here). Fourth irreducible wiring bump on this modal (969->989->993->998->1030); structural shrink tracked in #3501.",
     "_rebaseline_2026_07_19_7546_ghe_copilot_modal": "PR #7546 (GHE Copilot OAuth provider) own growth: OAuthModal.tsx 1030->1056 (gate units). Adds a gheUrl input state, routes ghe-copilot through the existing device-code branch, and threads gheUrl into the device-code request/poll extraData at the existing provider-switch chokepoints (+~24 lines, cohesive with the same pattern as #7399/#6636). The standalone GHE enterprise-URL config step JSX (originally +31 lines inline) was extracted to the new src/shared/components/oauthModal/GheConfigStep.tsx leaf component to minimize the bump; what remains is the irreducible provider-branch wiring. Fifth bump on this modal (969->989->993->998->1030->1056); structural shrink tracked in #3501.",
-    "_rebaseline_2026_07_19_7787_ic2_localdb_reexports": "PR #7787 (IC2 raw connections cache + lazy-decrypt) own growth: localDb.ts 805->807 (gate units, +2). localDb.ts is the re-export-only layer (hard rule #2 — no logic); the PR adds 4 new db/readCache re-exports (touchConnectionLastUsed, getCachedRawProviderConnections, getCachedProviderConnectionById, getCachedProviderNodes) required by existing barrel importers. Irreducible for a re-export list; frozen so it can only shrink.",
-    "_rebaseline_2026_07_20_7779_routingcombo_thread": "PR #7779 own growth: chatHelpers.ts 876->877 (+1, thread routingComboId into executeChatWithBreaker for compression-combo assignment). Frozen so it can only shrink.",
-    "_rebaseline_2026_07_20_7819_autocandidateoverrides_reexport": "PR for #7819 (Level 1+2: read-only auto/* candidate transparency + per-API-key exclusions) own growth: localDb.ts 807->808 (+1). Adds a single `export * from \"./db/autoCandidateOverrides\"` barrel re-export (hard rule #2 — no logic) for the new DB module backing per-apiKey candidate exclusions. Irreducible for a re-export list; frozen so it can only shrink.",
     "_rebaseline_2026_07_21_8027_grok_cli_auth_json_paste": "PR #8027 (RaviTharuma, fix(grok-cli) #7610) own growth: OAuthModal.tsx 1080->1100 (gate units). Requires the full ~/.grok/auth.json (with refresh_token) on the paste-import path instead of a bare JWT, at the existing paste-token chokepoint (renamed tab label, updated instructions/placeholder, textarea for the auth.json blob, inline error surface). The validation logic itself (parseGrokCliPasteToken, previously an inline ~75-line function) was extracted to the new src/lib/oauth/utils/grokCliAuthJson.ts leaf module — mirroring the #6636/#7546 extraction precedent — so only the irreducible UI wiring remains here. Sixth bump on this modal (969->989->993->998->1030->1056->1100); structural shrink tracked in #3501.",
+    "src/shared/components/OAuthModal.tsx": 1100,
+    "_rebaseline_2026_07_22_8213_requestloggerdetail_unblock_ui": "PR #8213 (hartmark, fix/gemini-tpm-quota-cooldown-wait) own growth: src/shared/components/RequestLoggerDetail.tsx 799->941 (+142, entirely this PR's diff — no other commit touched this file between the PR's merge-base and the release tip; crosses the general 800-line new-file cap so is frozen here for the first time). Adds a collapsible section header (open/expand-less toggle) plus per-log-entry unblock (`unblocking`/`cleared` state, isCombo503 detection) so the request-logger detail panel surfaces the same Gemini TPM cooldown-wait / model-lockout unblock action introduced by this PR at the individual-request level (mirrors the health-page bulk unblock action added in the same PR). Covered by the PR's own dashboard/resilience test additions.",
+    "src/shared/components/RequestLoggerDetail.tsx": 941,
+    "src/shared/components/RequestLoggerV2.tsx": 1629,
+    "src/shared/components/analytics/charts.tsx": 1558,
+    "_rebaseline_2026_07_10_6318_omp_letta": "PR #6318 (@hamsa0x7, omp+letta CLI integrations) own growth: cliTools.ts (+53 = 2 registry entries incl. omp docsUrl) and cliRuntime.ts (+18 = runtime-detection wiring for the 2 new tools). Cohesive registry/wiring growth at the existing chokepoints; scope reduced from the original 5 tools (pi/codewhale/jcode shipped separately).",
+    "src/shared/constants/cliTools.ts": 916,
+    "src/shared/constants/pricing.ts": 1662,
+    "src/shared/constants/providers.ts": 3276,
+    "src/shared/constants/sidebarVisibility.ts": 1198,
+    "src/shared/services/cliRuntime.ts": 1128,
+    "src/shared/validation/schemas.ts": 2523,
+    "_rebaseline_2026_06_28_5275_correlation_id_extract": "Extraction of the safe CorrelationId subset of #5275 (hartmark) — request correlation id stored in call_logs (migration 109) and returned via the X-Correlation-Id response header, WITHOUT the combo/resilience or build/lazy-loading changes (those stay in #5275). Own growth: callLogs.ts 975->985 (correlation_id column on CallLogSummaryRow + read/map), usageHistory.ts 983->988 (correlationId metadata normalize), chat.ts 1575->1632 (withCorrelationId response wiring + combo-failure log carrying correlationId), chatHelpers.ts new 811 (withCorrelationId helper + reqId threading; was 791<cap pre-feature). Cohesive request/logging chokepoint wiring; structural shrink of chat.ts tracked in #3501.",
+    "_rebaseline_2026_07_09_6678_routing_strategy_9router": "#6678 (SeaXen) — 9router-parity Routing Strategy settings card + per-provider/combo sticky-round-robin override. Own growth: ProviderDetailPageClient.tsx 784->786 (single ProviderAccountRoutingCard mount + import), auth.ts 2448->2458 (providerStrategies override resolution: fallbackStrategy/stickyRoundRobinLimit per-provider cascade in getProviderCredentials). Both additive, zero unrelated refactor; new UI/logic lives in new files (ProviderAccountRoutingCard.tsx, RoutingStrategyCard.tsx, rrState.ts::resolveComboStickyRoundRobinLimit). chat.ts value below reflects the current release tip (grown by other concurrent PRs, e.g. #6640), not this PR own change.",
+    "_rebaseline_2026_07_22_8213_chat_abandoned_target_abort": "PR #8213 (hartmark, fix/gemini-tpm-quota-cooldown-wait) own growth: src/sse/handlers/chat.ts 1794->1860 (+66, measured against the PR's own merge-base — the release tip separately carries an unrelated -5 net shrink from #8013's antigravity callable-catalog alignment, which this PR's branch does not include and this entry does not cover). Adds resolveDispatchClientRawRequest(): merges a per-target modelAbortSignal into clientRawRequest.signal (via mergeAbortSignals) so a combo target abandoned by comboTargetTimeoutMs actually observes its own abort and reaches its cleanup path, instead of hanging forever inside withRateLimit/acquireAccountSemaphore and leaking a permanent 'pending' dashboard entry (live incident, log id 1784418258231-14961a). Also wires combo-exhausted rejection logging to capture request body + attempted models via the new rejectedRequestUsage helper. Irreducible additions at the existing chat dispatch chokepoint. Covered by the PR's own combo-config + integration test additions.",
+    "_rebaseline_2026_07_23_8127_grok_weekly_quota": "#8127 (@apoapostolov) own growth: src/sse/handlers/chat.ts 1861->1865 (+4) — weekly quota tracking for grok-web wires a quota-fetch hook at the existing dispatch chokepoint. Thin wiring mirroring adjacent provider-quota branches; not extractable. Covered by tests/unit/grok-quota-fetcher.test.ts.",
+    "src/sse/handlers/chat.ts": 1865,
+    "_rebaseline_2026_07_20_7779_routingcombo_thread": "PR #7779 own growth: chatHelpers.ts 876->877 (+1, thread routingComboId into executeChatWithBreaker for compression-combo assignment). Frozen so it can only shrink.",
+    "src/sse/handlers/chatHelpers.ts": 878,
+    "src/sse/services/auth.ts": 2475,
+    "open-sse/executors/default.ts": 890,
+    "open-sse/translator/request/openai-responses.ts": 902,
+    "open-sse/executors/kiro.ts": 944,
+    "open-sse/translator/request/openai-to-claude.ts": 823,
+    "tests/unit/account-fallback-service.test.ts": 1572,
+    "tests/unit/provider-validation-specialty.test.ts": 2980,
+    "open-sse/executors/huggingchat.ts": 813,
+    "_rebaseline_2026_07_01_v3843_release_5609": "Rebaseline v3.8.43 (PR #5609 release reconciliation). DRIFT dos 109 commits do ciclo: 8 god-files existentes cresceram (ApiManagerPageClient 2983->3017, combos/page 4594->4608, AddApiKeyModal 868->869, providerPageHelpers 974->996, chat.ts 1635->1647, auth.ts 2401->2403, batchProcessor 828->915, combo.ts 3368->3387) + 2 novos acima do cap (huggingchat.ts 813, tests web-cookie-providers-new 827) + 4 test files cresceram. Modularizacao deferida (blast-radius mid-release); congelado no estado atual p/ o proximo ciclo ratchetar daqui.",
+    "src/lib/providers/validation/webProvidersA.ts": 809,
+    "src/lib/tokenHealthCheck.ts": 832,
+    "_rebaseline_2026_07_09_6587_kiro_api_key_auth": "PR #6587 (@strangersp) own growth for Kiro long-lived API-key auth, merged onto v3.8.47 tip: openai-to-kiro.ts 890->912 (+22, auth-header selection for API-key-vs-OAuth-token connections), providerLimits.ts 998->1000 (+2, API-key auth-type branch), translator-openai-to-kiro.test.ts 1234->1257 (+23), providers-page-utils.test.ts 1109->1107 (net -2 after merging with parallel release drift; connectionMatchesProviderCard api_key coverage added), provider-validation-specialty.test.ts 2856->2980 (+124 net after merge with parallel release drift; this PR also removed the file's `@typescript-eslint/no-explicit-any` eslint-suppression entry by fixing all `any` usages, adding typed replacements). Cohesive additive feature growth, well tested; not extractable without splitting the existing chokepoints mid-merge.",
+    "_rebaseline_2026_07_19_7787_ic2_localdb_reexports": "PR #7787 (IC2 raw connections cache + lazy-decrypt) own growth: localDb.ts 805->807 (gate units, +2). localDb.ts is the re-export-only layer (hard rule #2 — no logic); the PR adds 4 new db/readCache re-exports (touchConnectionLastUsed, getCachedRawProviderConnections, getCachedProviderConnectionById, getCachedProviderNodes) required by existing barrel importers. Irreducible for a re-export list; frozen so it can only shrink.",
+    "_rebaseline_2026_07_20_7819_autocandidateoverrides_reexport": "PR for #7819 (Level 1+2: read-only auto/* candidate transparency + per-API-key exclusions) own growth: localDb.ts 807->808 (+1). Adds a single `export * from \"./db/autoCandidateOverrides\"` barrel re-export (hard rule #2 — no logic) for the new DB module backing per-apiKey candidate exclusions. Irreducible for a re-export list; frozen so it can only shrink.",
+    "src/lib/localDb.ts": 808,
+    "_rebaseline_2026_07_12_v3847_mergeprs_tail": "v3.8.47 /merge-prs tail (owner-approved): src/lib/localDb.ts NEW>800 (799->805, +6 re-exports countFreeProxies + recordFreeProxySyncErrors/clearFreeProxySyncErrors/getFreeProxySyncErrors + FreeProxySyncErrors type for #6909 free-pool relay-repair; re-export-only per Hard Rule #2, not extractable).",
     "_rebaseline_2026_07_21_8034_compression_exclusions_sidebar": "#8034 (compression exclusions dashboard tab) own growth: sections.ts 796->806 (+10, one new COMPRESSION_CONTEXT_GROUP sidebar item linking /dashboard/compression/exclusions). The file was already 796/800 before this PR (organic growth from prior sidebar entries), so a single new nav item pushed it 6 lines over cap. Freezing at 806 (cannot grow further); the sidebar item array is data, not extractable logic.",
-    "_rebaseline_2026_07_22_7936_namespace_roundtrip": "#7936 (@RCrushMe, Responses-Chat namespace round-trip identity seam) own growth: open-sse/translator/response/openai-responses.ts 1092->1125 (+33) and open-sse/utils/stream.ts 2814->2869 (+55) — threading the namespace-identity seam through the Responses↔Chat translation + stream paths so tool-call namespaces survive the round-trip. Cohesive translation/stream wiring at existing chokepoints, frozen at new size.",
-    "_rebaseline_2026_07_22_8010_codex_responses_engine": "PR #8010 (@JxnLexn) own growth: open-sse/mcp-server/schemas/tools.ts 1497->1505 (+8 = threading the new \"codex-responses\" literal into the compressionConfigureInput strategy/autoTriggerMode Zod enums and setCompressionEngineInput engine enum, mirroring the existing rtk/omniglyph enum entries; no new tool). open-sse/services/compression/strategySelector.ts 1043->1054 (+11 = one new `if (mode === \"codex-responses\")` dispatch branch in runCompression that delegates 100% to the new codexResponsesEngine.apply, mirroring the existing rtk single-mode dispatch, plus threading config.codexResponsesConfig.preserveToolNames into the shared adaptBodyForCompression call at the 3 existing call sites). src/lib/db/compression.ts (untracked, new-file cap 800) 794->845 (+51 = normalizeCodexResponsesConfig, mirroring the existing normalizeRtkConfig normalizer, plus registering \"codex-responses\" in the COMPRESSION_MODES/STACKED_PIPELINE_ENGINE_IDS/SINGLE_MODE_ENGINE sets and the getCompressionSettings load/save switch) — added to the baseline at its current size. All three are cohesive dispatch/normalizer wiring at existing chokepoints (mirroring the prior compression-mode rebaselines #6534/#6556), not extractable without hiding the mode-dispatch boundary. Covered by tests/unit/compression/codex-responses.test.ts (6) + omniglyph-registries.test.ts/types.test.ts (22, updated for the new mode).",
-    "_rebaseline_2026_07_22_8034_compression_exclusions_persistence": "#8034 (compression exclusions) own growth: src/lib/db/compression.ts 845->850 (+5 = threading the new compressionExclusions field through the existing getCompressionSettings/saveCompressionSettings load/save switch over the shared key_value compression namespace — no new table, no raw SQL). Mirrors the prior compression-field rebaselines (#8010 codex-responses normalizer at the same chokepoint); the load/save switch is a single dispatch boundary, not extractable without hiding it. Covered by the PR's 8 node:test + 3 vitest cases.",
-    "_rebaseline_2026_07_22_8050_model_lockout_exact_family": "#8050 (@AndrianBalanescu) own growth: accountFallback.ts 1864->1892 (+28) — exact-vs-family model-lockout scoping (getModelLockKey/isModelLocked/clearModelLock/getModelLockoutInfo) so an Antigravity 404 for one bare model no longer hijacks the whole family cooldown. Cohesive lockout logic; frozen at new size.",
+    "_rebaseline_2026_07_23_8219_cache_ttl_settings_sidebar": "#8219 (@oyi77) own growth: sections.ts 806->813 (+7) — configurable model-catalog cache-TTL settings adds a new sidebar nav entry + its visibility wiring. Sidebar item array is data, not extractable logic; frozen at new size.",
+    "src/shared/constants/sidebarVisibility/sections.ts": 813,
     "_rebaseline_2026_07_22_8056_headroom_minrows": "#8056 (@RaviTharuma, persist Headroom minRows) own growth: src/lib/db/compression.ts 850->866 (+16 HeadroomConfig+DEFAULT_HEADROOM_CONFIG+normalize/store in get/updateCompressionSettings) and open-sse/services/compression/strategySelector.ts 1054->1060 (+6 merge settings.headroom into stacked stepConfig). Cohesive settings-persistence + stacked-merge wiring at existing chokepoints, frozen at new size.",
     "_rebaseline_2026_07_22_8081_reasoning_placeholder_guard": "#8081 (@Dingding-leo) own growth: openai-responses.ts 1125->1137 (+12) restructuring the reasoning-placeholder guard so it skips only the empty content block and still emits finish_reason/tool_calls in the same chunk. Cohesive translator wiring; frozen at new size.",
-    "_rebaseline_2026_07_22_8210_openrouter_midstream_error": "PR #8210 (hartmark, fix/openrouter-midstream-error-surfacing) own growth: open-sse/translator/response/openai-responses.ts 1137->1163 (+26) measured on the merged tip (release 1137 + this PR own growth). Adds a single new branch inside openaiToOpenAIResponsesResponse() that detects an OpenRouter-style mid-stream aggregator error (HTTP 200 SSE chunk with empty choices + a top-level error object) and surfaces it as state.upstreamError instead of silently falling through to the no-op/awaitingTrailingUsage path, which previously masked the failure as a false empty-success completion and skipped combo fallback. Irreducible call-site addition at the existing chunk-dispatch chokepoint (mirrors the Gemini-to-OpenAI translator's #4177 precedent for the same class of upstream error surfacing). Note: this baseline entry does NOT cover the separate pre-existing +11 drift already on the release tip from #8081/#8162 (1125->1136, unrelated reasoning-placeholder-stripping fix merged after this PR branched) — that drift belongs to the maintainer's rebaseline, not this PR.",
-    "_rebaseline_2026_07_22_8211_gemini_malformed_tool_choice": "PR #8211 (hartmark, fix/gemini-malformed-function-call-tool-choice) own growth: open-sse/translator/response/gemini-to-openai.ts 771->821 (+50, entirely this PR's diff — no other commit touched this file between the PR's merge-base and the release tip). Adds MALFORMED_FUNCTION_CALL/UNEXPECTED_TOOL_CALL handling inside geminiToOpenAIResponse(): synthesizes a `malformed_tool_call` tool_calls entry so finish_reason normalizes to the standard \"tool_calls\" instead of an unrecognized raw enum value that OpenAI-compatible clients (e.g. OpenClaw) silently ignore, and always synthesizes (rather than skipping when a real tool call already exists) so a malformed attempt alongside a real one in the same turn is not silently discarded. Irreducible cohesive addition at the existing candidate/finishReason translation chokepoint (mirrors the 9router#2462 raw-finish-reason precedent immediately below it in the same function). Covered by the PR's own tests/unit test additions for both the malformed-only and malformed-plus-real-call cases.",
-    "_rebaseline_2026_07_22_8213_chat_abandoned_target_abort": "PR #8213 (hartmark, fix/gemini-tpm-quota-cooldown-wait) own growth: src/sse/handlers/chat.ts 1794->1860 (+66, measured against the PR's own merge-base — the release tip separately carries an unrelated -5 net shrink from #8013's antigravity callable-catalog alignment, which this PR's branch does not include and this entry does not cover). Adds resolveDispatchClientRawRequest(): merges a per-target modelAbortSignal into clientRawRequest.signal (via mergeAbortSignals) so a combo target abandoned by comboTargetTimeoutMs actually observes its own abort and reaches its cleanup path, instead of hanging forever inside withRateLimit/acquireAccountSemaphore and leaking a permanent 'pending' dashboard entry (live incident, log id 1784418258231-14961a). Also wires combo-exhausted rejection logging to capture request body + attempted models via the new rejectedRequestUsage helper. Irreducible additions at the existing chat dispatch chokepoint. Covered by the PR's own combo-config + integration test additions.",
-    "_rebaseline_2026_07_22_8213_combo_cooldown_wait_recording": "PR #8213 (hartmark, fix/gemini-tpm-quota-cooldown-wait) own growth: open-sse/services/combo.ts 3548->3604 (+56, entirely this PR's diff — no other commit touched this file between the PR's merge-base and the release tip). Fixes combo cooldown-wait state recording so a bogus 503 is no longer crystallized when the cooldown-wait vars reset every setTry, adds an OpenAI-format SSE error frame path for combo-exhausted rejections (capturing request body + attempted models), and gives an abandoned per-target dispatch its own timeout instead of leaking a permanent 'pending' dashboard entry. Irreducible additions at the existing handleComboChat dispatch/retry chokepoint (mirrors the prior quota-share/headroom/task-aware strategy-branch precedents already frozen in this file). Covered by the PR's own combo-config + Gemini TPM-ceiling benchmark test additions.",
-    "_rebaseline_2026_07_22_8213_gemini_tpm_quota_cooldown_wait": "PR #8213 (hartmark, fix/gemini-tpm-quota-cooldown-wait) own growth: open-sse/services/accountFallback.ts 1857->1932 on the merged tip (release 1892 incl #8050 +35, plus this PR own growth +40); measured against the PR own merge-base was 1857->1898 (+41 — the release tip separately carries an unrelated +34 from #8050's antigravity 404 model-not-found lockout scoping, which this PR's branch does not include and this entry does not cover). Own growth is the Gemini TPM-ceiling classification + cooldown-wait wiring feeding into the combo cooldown-wait state machine (rate-limit wedge recovery) introduced by this PR's commit series. Irreducible additions at the existing account-fallback/model-lockout chokepoint. Covered by the PR's own gemini-rate-limit-tracker and TPM-ceiling benchmark test additions.",
-    "_rebaseline_2026_07_22_8213_health_unblock_model_cooldowns": "PR #8213 (hartmark, fix/gemini-tpm-quota-cooldown-wait) own growth: src/app/(dashboard)/dashboard/health/page.tsx 1094->1165 (+71, entirely this PR's diff — no other commit touched this file between the PR's merge-base and the release tip). Adds handleUnblockAll/handleUnblockOne dashboard actions (DELETE /api/resilience/model-cooldowns) so an operator can manually clear a Gemini TPM-wedge model lockout surfaced by this PR's cooldown-wait fixes, instead of waiting out the ceiling. Irreducible UI wiring at the existing health-page action chokepoint. Covered by the PR's own dashboard/resilience test additions.",
-    "_rebaseline_2026_07_22_8213_requestloggerdetail_unblock_ui": "PR #8213 (hartmark, fix/gemini-tpm-quota-cooldown-wait) own growth: src/shared/components/RequestLoggerDetail.tsx 799->941 (+142, entirely this PR's diff — no other commit touched this file between the PR's merge-base and the release tip; crosses the general 800-line new-file cap so is frozen here for the first time). Adds a collapsible section header (open/expand-less toggle) plus per-log-entry unblock (`unblocking`/`cleared` state, isCombo503 detection) so the request-logger detail panel surfaces the same Gemini TPM cooldown-wait / model-lockout unblock action introduced by this PR at the individual-request level (mirrors the health-page bulk unblock action added in the same PR). Covered by the PR's own dashboard/resilience test additions.",
+    "open-sse/services/usage/antigravity.ts": 802,
     "_rebaseline_2026_07_22_fusion_8013_8098_antigravity": "Fusion of #8013 (backryun, catalog/IDE-CLI-split rewrite) + #8098 (nguyenha935, protocol-fidelity/fail-closed/credits/tool-cloaking): open-sse/services/usage/antigravity.ts NEW 802 (>cap 800, +2 — #8098 credits/tier usage service on #8013's profile-aware headers). Test growth (models-catalog-route 1605->1608, provider-models-route 1752->1757 from #8013 Gemini 3.6 catalog) tracked in testFrozen.",
-    "_rebaseline_2026_07_23_8127_grok_weekly_quota": "#8127 (@apoapostolov) own growth: src/sse/handlers/chat.ts 1861->1865 (+4) — weekly quota tracking for grok-web wires a quota-fetch hook at the existing dispatch chokepoint. Thin wiring mirroring adjacent provider-quota branches; not extractable. Covered by tests/unit/grok-quota-fetcher.test.ts.",
-    "_rebaseline_2026_07_23_8143_empty_catch_logging": "#8143 (@chirag127) own growth: open-sse/utils/stream.ts 2869->2887 (+18) — replacing empty catch blocks in the SSE stream subsystem with console.debug logging (Rule #6 silent-swallow fix, issues #8138-#8142). Cohesive logging additions at the existing catch chokepoints, not extractable; frozen at new size. Covered by tests/unit/stream-handler-catch-logging-8143.test.ts.",
-    "_rebaseline_2026_07_23_8219_cache_ttl_settings_sidebar": "#8219 (@oyi77) own growth: sections.ts 806->813 (+7) — configurable model-catalog cache-TTL settings adds a new sidebar nav entry + its visibility wiring. Sidebar item array is data, not extractable logic; frozen at new size.",
-    "_rebaseline_2026_07_23_8247_8248_model_unhealthy": "#8247+#8248 own growth: accountFallback.ts 1940->1941 (+1, irreducible import statement only — the substantive #8248 DEGRADED-pattern classifier was extracted into open-sse/config/errorConfig.ts, which has ample headroom, instead of growing this frozen file; #8247's fix is a single existing-line condition change, net zero lines). Scoping the credits-exhausted 403/429 branch to isCompatibleProvider() (per-model-quota openai/anthropic-compatible-* nicknames) so it stays model-scoped instead of terminalling the whole connection, and classifying NVIDIA NIM 'Function ... DEGRADED' 400 bodies as model-access-denied instead of a raw passthrough 400. Covered by tests/unit/8247-accountfallback-model-unhealthy.test.ts and tests/unit/8248-accountfallback-nvidia-degraded.test.ts.",
-    "_rebaseline_2026_07_23_8252_combo_400_advance": "#8252 (@RaviTharuma) own growth: accountFallback.ts 1932->1940 (+8) + combo.ts 3604->3630 (+26) — advance combo on model-scoped 400s wrapped as invalid/Bad-Request. Irreducible wiring at existing account-fallback + combo dispatch chokepoints. Covered by combo-model-scoped-400-advance.test.ts.",
-    "_rebaseline_2026_07_23_8266_alibaba_media": "#8266 (@backryun) own growth: imageRegistry.ts 821->979 (+158) — Alibaba-family media models (Qwen image/video, Bailian, Wan) added to the image/video registry. Registry model data, not extractable logic; frozen at new size.",
-    "_rebaseline_2026_07_24_8388_compression_detail_persist": "#8388 (compression engine DETAIL settings — Headroom/session-dedup/CCR — dropped on save) own growth: src/lib/db/compression.ts 866->872 (+6 = irreducible call-site wiring at the existing getCompressionSettings/updateCompressionSettings chokepoint: one import line, one `...buildDetailConfigDefaults()` spread in the seed config, and one `case \"sessionDedup\": case \"ccr\": applyDetailConfigUpdate(config, key, parsed); break;` load-switch case, mirroring the existing headroom/#8056 case immediately above it). The actual normalizer logic (normalizeSessionDedupConfig/normalizeCcrConfig, matching SESSION_DEDUP_SCHEMA/CCR_SCHEMA bounds) was EXTRACTED into a new leaf src/lib/db/compressionDetailNormalizers.ts (well under cap) so this frozen file only carries the minimal dispatch wiring. Covered by tests/unit/8388-compression-detail-persist.test.ts (schema-accept + full DB save->reload round-trip for both new sub-objects, plus a no-regression assertion on the existing headroom round-trip).",
-    "_rebaseline_2026_07_24_responses_toolcalls_log_summary": "hartmark, fix/responses-tool-calls-log-summary own growth: open-sse/translator/response/openai-responses.ts 1163->1174 (+11). closeToolCall() now also writes the completed tool call into the shared state.toolCalls Map (already populated by the openai-to-claude / claude-to-openai / gemini-to-openai response translators) so stream.ts's completion-log summary builder (which reads state.toolCalls, not this translator's own funcCallIds/funcNames/funcArgsBuf bookkeeping) reports finish_reason \"tool_calls\" and message.tool_calls for openai->openai-responses translated streams instead of always logging \"stop\" with no tool_calls — the actual client-facing SSE events were already correct; only the persisted call-log summary was wrong. Irreducible call-site addition at the existing tool-call-close chokepoint. Covered by the new regression test in tests/unit/translator-resp-openai-responses.test.ts.",
-    "_rebaseline_2026_07_25_8476_combo_input_bound_homogeneous_scope": "PR #8476 (herjarsa, fix/8375-8459-combo-image-fixes, #8375) own growth: open-sse/services/combo.ts 3642->3679 (+37 net: +29 the PR's own isInputBoundFailure short-circuit for deterministic context_length_exceeded/context_window_exceeded failures, +8 a /green-prs pre-merge fix scoping that short-circuit to homogeneous remainders only — the shipped code fired unconditionally on ANY target, regressing the intentional heterogeneous-combo fallback #6637/isContextOverflow400 protects, exactly as flagged by this PR's own review evidence but never actually implemented in the branch). The fix compares orderedTargets[i+1..] modelStr against the failing target's modelStr at the existing executeTarget dispatch chokepoint (mirrors the sameProviderNext precedent a few lines below) — irreducible call-site wiring, not extractable without hiding the dispatch boundary. Covered by tests/unit/combo-input-bound-failure-8375.test.ts (homogeneous pool still short-circuits) and the new tests/unit/combo-input-bound-heterogeneous-8375.test.ts (heterogeneous combo now correctly falls through to the larger-context target).",
-    "_rebaseline_2026_07_25_adobe_firefly_reference_images": "Follow-up to #8006: storage upload + referenceBlobs for image/video and /v1/images/edits dispatch. adobeFireflyClient.ts 1958->2317 (+upload helpers, extract sources, resolve blob ids). Note: 2317 not 2316 — check-file-size.mjs counts LOC via split(\"\\n\").length (counts the trailing-newline empty element), which is 1 higher than `wc -l` on a file ending in \\n; the PR's original entry (2316) was measured with wc -l and undercounted by 1 against the actual gate.",
-    "_rebaseline_pr1043_minimax_tts": "Upstream port decolua/9router#1043 (toanalien) own growth: audioSpeech.ts 965->1061 (+96). Adds MiniMax T2A v2 TTS dispatch (handleMinimaxSpeech + hexToBytes helper) — provider entry was already in audioRegistry (format: minimax-tts) but no handler existed, falling through to the OpenAI-compatible default that fails (T2A has custom shape + hex-encoded audio + base_resp envelope). New branch sits next to the other inline provider branches (xiaomi-mimo, coqui, tortoise, aws-polly) — extracting would just create indirection. Covered by tests/unit/minimax-tts-1043.test.ts (3 tests, GREEN: success, base_resp error, invalid-hex).",
-    "_rebaseline_pr4592_exclude_exhausted_auto": "Reconcile #4592 already-merged growth: combo.ts 2991->3036 (+45, terminal-status quota-cutoff exclusion in buildAutoCandidates + opt-in gate). Fast-gate PR->release does not run check:file-size.",
-    "open-sse/executors/antigravity.ts": 1528,
-    "open-sse/executors/base.ts": 1640,
-    "open-sse/executors/chatgpt-web.ts": 3241,
-    "open-sse/executors/codex.ts": 1562,
-    "open-sse/executors/cursor.ts": 1563,
-    "open-sse/executors/deepseek-web.ts": 1148,
-    "open-sse/executors/grok-web.ts": 1044,
-    "open-sse/executors/muse-spark-web.ts": 1405,
-    "open-sse/handlers/chatCore.ts": 5034,
-    "open-sse/handlers/imageGeneration.ts": 3101,
-    "open-sse/handlers/responseSanitizer.ts": 1128,
-    "open-sse/handlers/search.ts": 1536,
-    "open-sse/handlers/videoGeneration.ts": 1063,
-    "open-sse/mcp-server/schemas/tools.ts": 1505,
-    "open-sse/mcp-server/server.ts": 1411,
-    "open-sse/mcp-server/tools/advancedTools.ts": 1120,
-    "open-sse/services/accountFallback.ts": 1972,
-    "open-sse/services/adobeFireflyClient.ts": 2385,
-    "open-sse/services/claudeCodeCompatible.ts": 1202,
-    "open-sse/services/combo.ts": 3648,
-    "open-sse/services/compression/strategySelector.ts": 1060,
-    "open-sse/services/rateLimitManager.ts": 1167,
-    "open-sse/translator/response/openai-responses.ts": 1204,
-    "open-sse/utils/cursorAgentProtobuf.ts": 1505,
-    "open-sse/utils/stream.ts": 2889,
-    "src/app/(dashboard)/dashboard/HomePageClient.tsx": 1388,
-    "src/app/(dashboard)/dashboard/analytics/ComboHealthTab.tsx": 1031,
-    "src/app/(dashboard)/dashboard/api-manager/ApiManagerPageClient.tsx": 3117,
-    "src/app/(dashboard)/dashboard/cache/media/MediaPageClient.tsx": 1067,
-    "src/app/(dashboard)/dashboard/combos/page.tsx": 4647,
-    "src/app/(dashboard)/dashboard/costs/CostOverviewTab.tsx": 1283,
-    "src/app/(dashboard)/dashboard/costs/quota-share/components/PoolWizard.tsx": 1022,
-    "src/app/(dashboard)/dashboard/endpoint/EndpointPageClient.tsx": 2615,
-    "src/app/(dashboard)/dashboard/health/page.tsx": 1165,
-    "src/app/(dashboard)/dashboard/providers/[id]/components/modals/EditConnectionModal.tsx": 1316,
-    "src/app/(dashboard)/dashboard/providers/page.tsx": 1944,
-    "src/app/(dashboard)/dashboard/runtime/RuntimePageClient.tsx": 1201,
-    "src/app/(dashboard)/dashboard/settings/components/PricingTab.tsx": 1019,
-    "src/app/(dashboard)/dashboard/settings/components/ProxyRegistryManager.tsx": 1464,
-    "src/app/(dashboard)/dashboard/settings/components/ResilienceTab.tsx": 1123,
-    "src/app/(dashboard)/dashboard/settings/components/RoutingTab.tsx": 1629,
-    "src/app/(dashboard)/dashboard/settings/components/SystemStorageTab.tsx": 1573,
-    "src/app/(dashboard)/dashboard/usage/components/BudgetTab.tsx": 1028,
-    "src/app/(dashboard)/dashboard/usage/components/EvalsTab.tsx": 2148,
-    "src/app/(dashboard)/dashboard/usage/components/ProviderLimits/index.tsx": 1119,
-    "src/app/api/providers/[id]/models/route.ts": 2250,
-    "src/app/api/v1/models/catalog.ts": 1549,
-    "src/lib/tokenHealthCheck.ts": 1021,
-    "src/lib/db/apiKeys.ts": 1529,
-    "src/lib/db/core.ts": 1637,
-    "src/lib/db/migrationRunner.ts": 1084,
-    "src/lib/db/models.ts": 1097,
-    "src/lib/db/providers.ts": 1034,
-    "src/lib/memory/retrieval.ts": 1073,
-    "src/lib/tailscaleTunnel.ts": 1202,
-    "src/lib/usage/providerLimits.ts": 1013,
-    "src/shared/components/OAuthModal.tsx": 1134,
-    "src/shared/components/RequestLoggerV2.tsx": 1629,
-    "src/shared/components/analytics/charts.tsx": 1035,
-    "src/shared/services/cliRuntime.ts": 1122,
-    "src/sse/handlers/chat.ts": 1877,
-    "src/sse/services/auth.ts": 2508,
-    "tests/unit/account-fallback-service.test.ts": 1572,
-    "tests/unit/provider-validation-specialty.test.ts": 2985,
-    "open-sse/executors/hyperagent.ts": 1026,
-    "open-sse/executors/default.ts": 1042,
-    "open-sse/executors/kiro.ts": 1069
+    "_rebaseline_2026_07_22_8050_model_lockout_exact_family": "#8050 (@AndrianBalanescu) own growth: accountFallback.ts 1864->1892 (+28) — exact-vs-family model-lockout scoping (getModelLockKey/isModelLocked/clearModelLock/getModelLockoutInfo) so an Antigravity 404 for one bare model no longer hijacks the whole family cooldown. Cohesive lockout logic; frozen at new size."
   },
   "testCap": 1000,
   "testFrozen": {
@@ -326,7 +381,7 @@
     "tests/unit/stream-utils.test.ts": 2445,
     "tests/unit/token-refresh-service.test.ts": 1378,
     "tests/unit/translator-openai-responses-req.test.ts": 1194,
-    "tests/unit/translator-openai-to-gemini.test.ts": 1622,
+    "tests/unit/translator-openai-to-gemini.test.ts": 1619,
     "tests/unit/translator-openai-to-kiro.test.ts": 1275,
     "tests/unit/translator-resp-gemini-to-openai.test.ts": 1234,
     "tests/unit/usage-service-hardening.test.ts": 1483,
@@ -411,6 +466,132 @@
   "_rebaseline_2026_07_27_3850_relax_filesize_cap": "OWNER-APPROVED TEMPORARY relax for v3.8.50-3.8.54 PREPARE phase (docs/ROADMAP.md). cap 800->900 (+100), testCap 800->900 (+100). Targets: decompose-existing-frozen unchanged (frozen still only-shrink); this only relaxes the cap for NEW files in the decompose/extract-while-PREPARE phase (.51='executor registry in-place' and .52='combo.ts decomposition' create new leaf modules above 800). RE-TIGHTENING MANDATORY in v3.8.51: cap target 850 = 850 once decomposition wave stabilizes. SUPERSEDED by _rebaseline_2026_07_27_3850_relax_filesize_cap_v2_20pct (v1 +20% buffer) — retained for audit. Tracked via same roadmap issue.",
   "_rebaseline_2026_07_27_v3849_train1h": "Merge-train 1H (31 PRs) — owner-approved 2026-07-27. Two distinct causes, kept separate on purpose: (1) GENUINE irreducible growth at existing chokepoints — providerLimits/auth (#8632 Kimi quota-reset recovery), rateLimitManager (#8616 idle wedged limiters), models-catalog-route.test (#8610 OpenCode Go effort aliases); (2) COLLISION with #8585, which banked shrinks measured on the pre-train release tip while 30 sibling PRs in the SAME train grew those files again — chat/accountFallback (#8628), chatCore (#8613), videoGeneration (#8581), imageGeneration. The zero-headroom frozen entries cannot absorb either. Ceilings re-pinned to the post-merge tip; #8612 (also in this train) automates shrink-banking so this self-inflicted drift stops recurring. Detail: src/lib/usage/providerLimits.ts 1006->1013 (#8632); src/sse/services/auth.ts 2492->2508 (#8632); open-sse/services/rateLimitManager.ts 1014->1060 (#8616); src/sse/handlers/chat.ts 1842->1845 (#8628); open-sse/handlers/chatCore.ts 4939->4955 (#8613); open-sse/handlers/imageGeneration.ts 3100->3101 ((sem PR — teto do #8585)); open-sse/handlers/videoGeneration.ts 1038->1063 (#8581); open-sse/services/accountFallback.ts 1965->1966 (#8628); tests/unit/models-catalog-route.test.ts 1608->1636 (#8610)",
   "_rebaseline_2026_08_02_9242_token_health_transient": "PR #9242 (fix/refresh-circuit-transient): src/lib/tokenHealthCheck.ts 1021 (new file, above cap 1000). The file consolidates token-refresh health checking logic that was previously scattered across auth.ts and tokenRefresh.ts. Cohesive single-responsibility module for refresh circuit state management; not extractable without splitting the refresh state machine. Covered by tests/unit/tokenHealthCheck-transient.test.ts.",
+  "frozen": {
+    "_rebaseline_2026_06_22_4644_deepseek_web_tools": "PR #4644 (BugsBag/robust deepseek-web tool-call parsing): open-sse/executors/deepseek-web.ts 1117->1125 (+8). The new agentic tool-call path emits surrounding text + reasoning before tool_calls and swaps to the dedicated deepseekWebTools.ts parser; the +8 lines are cohesive wiring at the existing transformSSE chokepoint (the parser itself lives in the new deepseekWebTools.ts file, already under cap). The PR's own fast-gate (PR->release) does not run check:file-size, so this surfaced only at release reconcile. Covered by tests/unit/deepseek-web-tools-variants.test.ts + deepseek-web-tools-execute.test.ts.",
+    "_rebaseline_2026_06_23_4712_deepseek_web_tool_results": "PR for #4712 (deepseek-web drops role:tool): open-sse/executors/deepseek-web.ts 1125->1148 (+23). messagesToPrompt() now folds role:\"tool\" results into the single-prompt transcript (recovering the tool name from the preceding assistant tool_calls by tool_call_id) instead of silently dropping them; the lines are cohesive wiring inside the existing function. Covered by tests/unit/deepseek-web-tool-result-prompt-4712.test.ts.",
+    "_rebaseline_2026_06_24_headroom_strategy": "Headroom-aware connection selection (dario technique): combo.ts 3168->3180 (+12 = a new `else if (strategy === \"headroom\")` dispatch branch in handleComboChat that delegates to orderTargetsByHeadroom + its log line, plus the import). The actual logic lives OUT of the god-file: the pure ranker rankByHeadroom/computeHeadroom is the new leaf open-sse/services/combo/headroomRanking.ts (91 LOC, <cap) and the async orderer orderTargetsByHeadroom is appended to the existing open-sse/services/combo/quotaStrategies.ts (<cap) next to its sibling reset-aware/reset-window orderers (reuses their connection-expansion machinery). headroom = 1 - max(util_5h, util_7d) from getSaturation (src/lib/quota/saturationSignals.ts), prefers the connection with the most free capacity. Only the dispatch wiring is irreducible at the existing combo strategy chokepoint (mirrors the reset-aware/reset-window/context-optimized branches); not extractable without hiding the call site. fill-first stays default; all existing strategies untouched. Covered by tests/unit/combo-headroom-ranking.test.ts (pure helper) + tests/unit/combo-headroom-strategy.test.ts (orderer, saturation injected). Structural shrink of combo.ts tracked in #3501.",
+    "_rebaseline_2026_06_24_quota_share_strategy": "Dedicated quota-share strategy (Phase 3 #9): combo.ts 3180->3190 (+10 = one new `else if (strategy === \"quota-share\")` dispatch branch in handleComboChat that delegates 100% to selectQuotaShareTarget + its log line, plus the import). All the new logic lives OUT of the god-file in two new leaves under open-sse/services/combo/: quotaShareInflight.ts (in-flight counter with TTL/lease, ~150 LOC <cap) and quotaShareStrategy.ts (per-model bucket gating via isBucketSaturated + DRR proportional to weight + P2C over in-flight, ~240 LOC <cap). Only the dispatch wiring is irreducible at the existing combo strategy chokepoint (mirrors the headroom/reset-aware/reset-window/context-optimized branches); not extractable without hiding the call site. ZERO existing strategy cases were modified — only this branch was added, and the qtSd/ combos switched from fill-first to quota-share in src/lib/quota/quotaCombos.ts. Covered by tests/unit/quota-share-strategy.test.ts (gating, DRR fairness, P2C in-flight, fail-open, activation). Structural shrink of combo.ts tracked in #3501.",
+    "_rebaseline_2026_06_24_task_aware_routing": "Task-aware routing strategy (port PR #2045, OmniRoute #4945): combo.ts 3190->3225 (+35) = one new `else if (strategy === \"task-aware\")` dispatch branch delegating 100% to selectTaskAwareTarget + its imports/log lines. All scoring/classification logic lives OUT of the god-file in the new leaf open-sse/services/taskAwareRouting.ts (553 LOC <cap). Only the dispatch wiring is irreducible at the existing combo strategy chokepoint (mirrors quota-share/headroom/reset-aware branches). ZERO existing strategy cases modified. Covered by tests/unit/combo-task-aware.test.ts (35 tests). Structural shrink of combo.ts tracked in #3501.",
+    "_rebaseline_2026_06_26_fidelity_gate_extraction": "Milestone-B fidelity-gate wiring residual: bodyToText+gateAdvance extracted to fidelityGateStep.ts (889->854, -35), but the StackOptions.fidelityGate field, the `const fidelityGate` reads at the two stacked-loop dispatch chokepoints, and the import of FidelityGateConfig are irreducible wiring that cannot leave strategySelector without an architectural refactor of the pre-existing stacked pipeline. Net: 889->854 (+6 vs the pre-Milestone-B frozen 848). Covered by tests/unit/compression/*.test.ts (940 pass).",
+    "_rebaseline_2026_06_27_5193_5203_antigravity_oauthmodal": "Antigravity remote-login own growth: OAuthModal.tsx 960->969 (gate units). #5193 (+~4: remote paste instruction shown for all remote incl. Google + its rationale comment) and #5203 (+~5: handleManualSubmit credential-blob branch + button guard; submit logic extracted to oauthBlobSubmit.ts to minimize). Frozen set to the SUM so either merge order passes. Cohesive at the existing manual-submit chokepoint.",
+    "_rebaseline_2026_06_27_5193_antigravity_basered": "Base-red (pre-existing release drift, fast-gate PR->release skips check:file-size): accountFallback.ts 1773->1777 and src/app/api/providers/[id]/test/route.ts 924->940 were already over their frozen caps on release/v3.8.39 independent of any antigravity change. Owner chose to rebaseline (keep the documented issue-reference comments #1846/#1449/#347 etc.) rather than accept the contributor comment-stripping in #5200/#5198. Reverted #5200 to restore the comments; bumped these two frozen caps to the actual base sizes. No logic change.",
+    "_rebaseline_2026_06_28_5237_impersonation_ua_refresh": "PR #5237 (refresh impersonation UAs): grok-web.ts 1871->1873 (+2), muse-spark-web.ts 1284->1302 (+18), perplexity-web.ts 1013->1032 (+19). Net semantic change in each file is a single User-Agent constant (Chrome 147->149 for grok/muse; perplexity kept at Firefox 148 to stay matched with the firefox_148 TLS profile — the contributor's 152 bump was reverted to avoid a UA-vs-JA3 mismatch, #2459). The growth is Prettier reflow that lint-staged unavoidably applies to these grandfathered long-line files the moment they are touched; not extractable. src/sse/services/auth.ts 2336->2401 in the same reconcile is #5222's antigravity-LRU-retry growth that merged via --admin without a baseline bump.",
+    "_rebaseline_2026_06_28_5243_risk_gate_prepass": "PR #5243 (compression risk-gate pre-pass) own growth: open-sse/services/compression/strategySelector.ts 854->899 (+45). The three exported entry points (applyCompression/applyStackedCompression/applyStackedCompressionAsync) become thin wrappers over pure-extracted private bodies (runCompression/runStackedCompression/runStackedCompressionAsync) so the risk-gate mask->run->restore wrapper sits strictly OUTSIDE the per-step loop — a single universal integration point. The wrapper logic itself (resolveRiskGate/withRiskGate) lives in the new riskGate/strategyWrap.ts (<cap); the residual growth is the duplicated thin-wrapper signatures + the extracted bodies' dispatch boundary, guarded by a byte-identical parity test (riskGateIntegration). Default off (DEFAULT_COMPRESSION_CONFIG unchanged). Not extractable without hiding the dispatch boundary, mirroring prior compression rebaselines. Structural shrink tracked in #3501.",
+    "_rebaseline_2026_06_28_5275_correlation_id_extract": "Extraction of the safe CorrelationId subset of #5275 (hartmark) — request correlation id stored in call_logs (migration 109) and returned via the X-Correlation-Id response header, WITHOUT the combo/resilience or build/lazy-loading changes (those stay in #5275). Own growth: callLogs.ts 975->985 (correlation_id column on CallLogSummaryRow + read/map), usageHistory.ts 983->988 (correlationId metadata normalize), chat.ts 1575->1632 (withCorrelationId response wiring + combo-failure log carrying correlationId), chatHelpers.ts new 811 (withCorrelationId helper + reqId threading; was 791<cap pre-feature). Cohesive request/logging chokepoint wiring; structural shrink of chat.ts tracked in #3501.",
+    "_rebaseline_2026_06_29_4038_cas_guard": "PR (#4038) own growth: tokenRefresh.ts 2103->2181 (+78 = the compare-and-swap guard on the refresh persist — runWithCasGuard/getActiveCasGuard AsyncLocalStorage pair mirroring runWithOnPersist, casGuardShouldSkipPersist that rereads the row right before persisting and skips the write when a concurrent writer already rotated the refresh_token past the one presented, plus getCasGuardStats counters). Fixes the sibling-rotation-revert → token-family-revocation storm. Gated behind an active guard (opt-in; no guard => byte-identical). Wiring lives at the two persist chokepoints inside getAccessToken; the comparison reuses wasRefreshTokenRotated from refreshSerializer. Not extractable without splitting the refresh hot path.",
+    "_rebaseline_2026_06_29_5286_memoization": "PR #5286 own growth: strategySelector.ts 899->960 (+61 = the opt-in result-memoization branches in applyCompression/applyCompressionAsync — principal+determinism gate, makeMemoKey lookup/store with model+supportsVision folded into the key, recompute-with-memo-off). Default off (memoizeCompressionResults), so zero behavior change. The memo helpers live in the leaf resultMemo.ts (<cap); the chokepoint wiring here is not extractable. Structural shrink of this hot-path file tracked in #3501.",
+    "_rebaseline_2026_07_01_v3843_release_5609": "Rebaseline v3.8.43 (PR #5609 release reconciliation). DRIFT dos 109 commits do ciclo: 8 god-files existentes cresceram (ApiManagerPageClient 2983->3017, combos/page 4594->4608, AddApiKeyModal 868->869, providerPageHelpers 974->996, chat.ts 1635->1647, auth.ts 2401->2403, batchProcessor 828->915, combo.ts 3368->3387) + 2 novos acima do cap (huggingchat.ts 813, tests web-cookie-providers-new 827) + 4 test files cresceram. Modularizacao deferida (blast-radius mid-release); congelado no estado atual p/ o proximo ciclo ratchetar daqui.",
+    "_rebaseline_2026_07_02_5816_qoder": "PR #5816 (@AgentKiller45, qoder PAT via qodercli): qoderCli.ts 666->989, new-above-cap frozen (owner-approved baseline freeze). The growth is the legitimate PAT job-token exchange + quota parsing CLI transport (the pure-JS Cosy path 500'd on every PAT request); extracting the spawn/parse helpers now would just add indirection to a contributor PR mid-merge. Test frozen also raised for this PR's coverage growth: providers-page-utils.test.ts 1052->1092. Additionally clears an inherited base-red from the already-merged #5933 (codex json_schema->text.format): translator-openai-responses-req.test.ts 1097->1172 (+75 regression tests, no offending branch left). All remain frozen (cannot grow further); release captain's rebaseline-at-release supersedes.",
+    "_rebaseline_2026_07_09_6126_clinepass_dual_auth": "PR #6126 (@hajilok, dual-auth ClinePass) own growth: tokenRefresh.ts 2181->2182 (+1 = a single `case \"clinepass\":` fallthrough label added to the existing `case \"cline\":` in _getAccessTokenInternal's provider switch, so clinepass token refresh dispatches to the already-shared refreshClineToken() instead of silently falling through to the generic OAuth refresh). Irreducible 1-line switch-case wiring at the existing chokepoint; the header-building logic for the same feature was extracted to a new leaf src/shared/utils/clineAuth.ts::buildClinepassHeaders() (well under cap) to avoid growing open-sse/executors/default.ts. Covered by tests/unit/clinepass-provider.test.ts.",
+    "_rebaseline_2026_07_09_6363_kiro_external_idp": "PR #6363 (@artickc, Kiro external IdP) own growth: tokenRefresh.ts 2182->2249 (+67 = the external_idp refresh branch inside refreshKiroToken — standard public-client OAuth2 refresh_token grant against the org IdP tokenEndpoint via buildExternalIdpRefreshParams/isExternalIdpAuthMethod from the new leaf open-sse/services/kiroExternalIdp.ts, with invalid_grant/invalid_client -> unrecoverable_refresh_error mapping). Cohesive addition at the existing refreshKiroToken chokepoint. Covered by tests/unit/kiro-external-idp.test.ts.",
+    "_rebaseline_2026_07_09_6587_kiro_api_key_auth": "PR #6587 (@strangersp) own growth for Kiro long-lived API-key auth, merged onto v3.8.47 tip: openai-to-kiro.ts 890->912 (+22, auth-header selection for API-key-vs-OAuth-token connections), providerLimits.ts 998->1000 (+2, API-key auth-type branch), translator-openai-to-kiro.test.ts 1234->1257 (+23), providers-page-utils.test.ts 1109->1107 (net -2 after merging with parallel release drift; connectionMatchesProviderCard api_key coverage added), provider-validation-specialty.test.ts 2856->2980 (+124 net after merge with parallel release drift; this PR also removed the file's `@typescript-eslint/no-explicit-any` eslint-suppression entry by fixing all `any` usages, adding typed replacements). Cohesive additive feature growth, well tested; not extractable without splitting the existing chokepoints mid-merge.",
+    "_rebaseline_2026_07_09_6678_routing_strategy_9router": "#6678 (SeaXen) — 9router-parity Routing Strategy settings card + per-provider/combo sticky-round-robin override. Own growth: ProviderDetailPageClient.tsx 784->786 (single ProviderAccountRoutingCard mount + import), auth.ts 2448->2458 (providerStrategies override resolution: fallbackStrategy/stickyRoundRobinLimit per-provider cascade in getProviderCredentials). Both additive, zero unrelated refactor; new UI/logic lives in new files (ProviderAccountRoutingCard.tsx, RoutingStrategyCard.tsx, rrState.ts::resolveComboStickyRoundRobinLimit). chat.ts value below reflects the current release tip (grown by other concurrent PRs, e.g. #6640), not this PR own change.",
+    "_rebaseline_2026_07_10_6318_omp_letta": "PR #6318 (@hamsa0x7, omp+letta CLI integrations) own growth: cliTools.ts (+53 = 2 registry entries incl. omp docsUrl) and cliRuntime.ts (+18 = runtime-detection wiring for the 2 new tools). Cohesive registry/wiring growth at the existing chokepoints; scope reduced from the original 5 tools (pi/codewhale/jcode shipped separately).",
+    "_rebaseline_2026_07_10_gcf_v3_2_decode": "PR #6838 own growth: new vendored file open-sse/services/compression/engines/headroom/gcf/decode_generic.ts frozen at 880 (> 800 cap). It is the vendored GCF generic-profile decoder (spec v3.2 nested flattening plus the prototype-pollution / hasOwnProperty hardening added in this PR's Gemini review). Kept as one file faithful to upstream gcf-typescript so re-vendoring stays a clean copy rather than a re-split each cycle (sibling generic.ts/scalar.ts stay < cap; extraction would also fragment the file's frozen eslint no-explicit-any suppressions). Round-trip + prototype-pollution regression coverage in tests/unit/compression/headroom-smartcrusher.test.ts. Frozen: only shrinks from here.",
+    "_rebaseline_2026_07_12_v3847_mergeprs_tail": "v3.8.47 /merge-prs tail (owner-approved): src/lib/localDb.ts NEW>800 (799->805, +6 re-exports countFreeProxies + recordFreeProxySyncErrors/clearFreeProxySyncErrors/getFreeProxySyncErrors + FreeProxySyncErrors type for #6909 free-pool relay-repair; re-export-only per Hard Rule #2, not extractable).",
+    "_rebaseline_2026_07_15_7070_combos_memo": "PR #7070 (perf/p1-memo) own growth: src/app/(dashboard)/dashboard/combos/page.tsx 4655->4656 (+1 = React.memo wrapping of ComboCard). Covered by tests/unit/ui/combos-page-smoke.test.tsx.",
+    "_rebaseline_2026_07_18_7399_xai_oauth_modal": "PR #7399 (xAI OAuth PKCE) own growth: OAuthModal.tsx 993->998 (+5 = provider entry + PKCE flow branch wiring at the existing provider-switch chokepoint; the provider logic itself lives in src/lib/oauth/providers/xai-oauth.ts, new leaf). Third irreducible wiring bump on this modal (969->989->993->998); structural shrink tracked in #3501.",
+    "_rebaseline_2026_07_19_6636_codex_session_json": "#6636 own growth: OAuthModal.tsx 998->1030 (gate units, split(\"\\n\").length incl. trailing newline; +32 = session-JSON paste branch for handleManualSubmit plus a shared submitCodexAccessToken() helper extracted from the pre-existing bare-JWT branch, mirroring the #5203 oauthBlobSubmit.ts extraction precedent; the normalizer logic itself lives in the new src/lib/oauth/utils/codexSessionImport.ts leaf module, not here). Fourth irreducible wiring bump on this modal (969->989->993->998->1030); structural shrink tracked in #3501.",
+    "_rebaseline_2026_07_19_7546_ghe_copilot_modal": "PR #7546 (GHE Copilot OAuth provider) own growth: OAuthModal.tsx 1030->1056 (gate units). Adds a gheUrl input state, routes ghe-copilot through the existing device-code branch, and threads gheUrl into the device-code request/poll extraData at the existing provider-switch chokepoints (+~24 lines, cohesive with the same pattern as #7399/#6636). The standalone GHE enterprise-URL config step JSX (originally +31 lines inline) was extracted to the new src/shared/components/oauthModal/GheConfigStep.tsx leaf component to minimize the bump; what remains is the irreducible provider-branch wiring. Fifth bump on this modal (969->989->993->998->1030->1056); structural shrink tracked in #3501.",
+    "_rebaseline_2026_07_19_7787_ic2_localdb_reexports": "PR #7787 (IC2 raw connections cache + lazy-decrypt) own growth: localDb.ts 805->807 (gate units, +2). localDb.ts is the re-export-only layer (hard rule #2 — no logic); the PR adds 4 new db/readCache re-exports (touchConnectionLastUsed, getCachedRawProviderConnections, getCachedProviderConnectionById, getCachedProviderNodes) required by existing barrel importers. Irreducible for a re-export list; frozen so it can only shrink.",
+    "_rebaseline_2026_07_20_7779_routingcombo_thread": "PR #7779 own growth: chatHelpers.ts 876->877 (+1, thread routingComboId into executeChatWithBreaker for compression-combo assignment). Frozen so it can only shrink.",
+    "_rebaseline_2026_07_20_7819_autocandidateoverrides_reexport": "PR for #7819 (Level 1+2: read-only auto/* candidate transparency + per-API-key exclusions) own growth: localDb.ts 807->808 (+1). Adds a single `export * from \"./db/autoCandidateOverrides\"` barrel re-export (hard rule #2 — no logic) for the new DB module backing per-apiKey candidate exclusions. Irreducible for a re-export list; frozen so it can only shrink.",
+    "_rebaseline_2026_07_21_8027_grok_cli_auth_json_paste": "PR #8027 (RaviTharuma, fix(grok-cli) #7610) own growth: OAuthModal.tsx 1080->1100 (gate units). Requires the full ~/.grok/auth.json (with refresh_token) on the paste-import path instead of a bare JWT, at the existing paste-token chokepoint (renamed tab label, updated instructions/placeholder, textarea for the auth.json blob, inline error surface). The validation logic itself (parseGrokCliPasteToken, previously an inline ~75-line function) was extracted to the new src/lib/oauth/utils/grokCliAuthJson.ts leaf module — mirroring the #6636/#7546 extraction precedent — so only the irreducible UI wiring remains here. Sixth bump on this modal (969->989->993->998->1030->1056->1100); structural shrink tracked in #3501.",
+    "_rebaseline_2026_07_21_8034_compression_exclusions_sidebar": "#8034 (compression exclusions dashboard tab) own growth: sections.ts 796->806 (+10, one new COMPRESSION_CONTEXT_GROUP sidebar item linking /dashboard/compression/exclusions). The file was already 796/800 before this PR (organic growth from prior sidebar entries), so a single new nav item pushed it 6 lines over cap. Freezing at 806 (cannot grow further); the sidebar item array is data, not extractable logic.",
+    "_rebaseline_2026_07_22_7936_namespace_roundtrip": "#7936 (@RCrushMe, Responses-Chat namespace round-trip identity seam) own growth: open-sse/translator/response/openai-responses.ts 1092->1125 (+33) and open-sse/utils/stream.ts 2814->2869 (+55) — threading the namespace-identity seam through the Responses↔Chat translation + stream paths so tool-call namespaces survive the round-trip. Cohesive translation/stream wiring at existing chokepoints, frozen at new size.",
+    "_rebaseline_2026_07_22_8010_codex_responses_engine": "PR #8010 (@JxnLexn) own growth: open-sse/mcp-server/schemas/tools.ts 1497->1505 (+8 = threading the new \"codex-responses\" literal into the compressionConfigureInput strategy/autoTriggerMode Zod enums and setCompressionEngineInput engine enum, mirroring the existing rtk/omniglyph enum entries; no new tool). open-sse/services/compression/strategySelector.ts 1043->1054 (+11 = one new `if (mode === \"codex-responses\")` dispatch branch in runCompression that delegates 100% to the new codexResponsesEngine.apply, mirroring the existing rtk single-mode dispatch, plus threading config.codexResponsesConfig.preserveToolNames into the shared adaptBodyForCompression call at the 3 existing call sites). src/lib/db/compression.ts (untracked, new-file cap 800) 794->845 (+51 = normalizeCodexResponsesConfig, mirroring the existing normalizeRtkConfig normalizer, plus registering \"codex-responses\" in the COMPRESSION_MODES/STACKED_PIPELINE_ENGINE_IDS/SINGLE_MODE_ENGINE sets and the getCompressionSettings load/save switch) — added to the baseline at its current size. All three are cohesive dispatch/normalizer wiring at existing chokepoints (mirroring the prior compression-mode rebaselines #6534/#6556), not extractable without hiding the mode-dispatch boundary. Covered by tests/unit/compression/codex-responses.test.ts (6) + omniglyph-registries.test.ts/types.test.ts (22, updated for the new mode).",
+    "_rebaseline_2026_07_22_8034_compression_exclusions_persistence": "#8034 (compression exclusions) own growth: src/lib/db/compression.ts 845->850 (+5 = threading the new compressionExclusions field through the existing getCompressionSettings/saveCompressionSettings load/save switch over the shared key_value compression namespace — no new table, no raw SQL). Mirrors the prior compression-field rebaselines (#8010 codex-responses normalizer at the same chokepoint); the load/save switch is a single dispatch boundary, not extractable without hiding it. Covered by the PR's 8 node:test + 3 vitest cases.",
+    "_rebaseline_2026_07_22_8050_model_lockout_exact_family": "#8050 (@AndrianBalanescu) own growth: accountFallback.ts 1864->1892 (+28) — exact-vs-family model-lockout scoping (getModelLockKey/isModelLocked/clearModelLock/getModelLockoutInfo) so an Antigravity 404 for one bare model no longer hijacks the whole family cooldown. Cohesive lockout logic; frozen at new size.",
+    "_rebaseline_2026_07_22_8056_headroom_minrows": "#8056 (@RaviTharuma, persist Headroom minRows) own growth: src/lib/db/compression.ts 850->866 (+16 HeadroomConfig+DEFAULT_HEADROOM_CONFIG+normalize/store in get/updateCompressionSettings) and open-sse/services/compression/strategySelector.ts 1054->1060 (+6 merge settings.headroom into stacked stepConfig). Cohesive settings-persistence + stacked-merge wiring at existing chokepoints, frozen at new size.",
+    "_rebaseline_2026_07_22_8081_reasoning_placeholder_guard": "#8081 (@Dingding-leo) own growth: openai-responses.ts 1125->1137 (+12) restructuring the reasoning-placeholder guard so it skips only the empty content block and still emits finish_reason/tool_calls in the same chunk. Cohesive translator wiring; frozen at new size.",
+    "_rebaseline_2026_07_22_8210_openrouter_midstream_error": "PR #8210 (hartmark, fix/openrouter-midstream-error-surfacing) own growth: open-sse/translator/response/openai-responses.ts 1137->1163 (+26) measured on the merged tip (release 1137 + this PR own growth). Adds a single new branch inside openaiToOpenAIResponsesResponse() that detects an OpenRouter-style mid-stream aggregator error (HTTP 200 SSE chunk with empty choices + a top-level error object) and surfaces it as state.upstreamError instead of silently falling through to the no-op/awaitingTrailingUsage path, which previously masked the failure as a false empty-success completion and skipped combo fallback. Irreducible call-site addition at the existing chunk-dispatch chokepoint (mirrors the Gemini-to-OpenAI translator's #4177 precedent for the same class of upstream error surfacing). Note: this baseline entry does NOT cover the separate pre-existing +11 drift already on the release tip from #8081/#8162 (1125->1136, unrelated reasoning-placeholder-stripping fix merged after this PR branched) — that drift belongs to the maintainer's rebaseline, not this PR.",
+    "_rebaseline_2026_07_22_8211_gemini_malformed_tool_choice": "PR #8211 (hartmark, fix/gemini-malformed-function-call-tool-choice) own growth: open-sse/translator/response/gemini-to-openai.ts 771->821 (+50, entirely this PR's diff — no other commit touched this file between the PR's merge-base and the release tip). Adds MALFORMED_FUNCTION_CALL/UNEXPECTED_TOOL_CALL handling inside geminiToOpenAIResponse(): synthesizes a `malformed_tool_call` tool_calls entry so finish_reason normalizes to the standard \"tool_calls\" instead of an unrecognized raw enum value that OpenAI-compatible clients (e.g. OpenClaw) silently ignore, and always synthesizes (rather than skipping when a real tool call already exists) so a malformed attempt alongside a real one in the same turn is not silently discarded. Irreducible cohesive addition at the existing candidate/finishReason translation chokepoint (mirrors the 9router#2462 raw-finish-reason precedent immediately below it in the same function). Covered by the PR's own tests/unit test additions for both the malformed-only and malformed-plus-real-call cases.",
+    "_rebaseline_2026_07_22_8213_chat_abandoned_target_abort": "PR #8213 (hartmark, fix/gemini-tpm-quota-cooldown-wait) own growth: src/sse/handlers/chat.ts 1794->1860 (+66, measured against the PR's own merge-base — the release tip separately carries an unrelated -5 net shrink from #8013's antigravity callable-catalog alignment, which this PR's branch does not include and this entry does not cover). Adds resolveDispatchClientRawRequest(): merges a per-target modelAbortSignal into clientRawRequest.signal (via mergeAbortSignals) so a combo target abandoned by comboTargetTimeoutMs actually observes its own abort and reaches its cleanup path, instead of hanging forever inside withRateLimit/acquireAccountSemaphore and leaking a permanent 'pending' dashboard entry (live incident, log id 1784418258231-14961a). Also wires combo-exhausted rejection logging to capture request body + attempted models via the new rejectedRequestUsage helper. Irreducible additions at the existing chat dispatch chokepoint. Covered by the PR's own combo-config + integration test additions.",
+    "_rebaseline_2026_07_22_8213_combo_cooldown_wait_recording": "PR #8213 (hartmark, fix/gemini-tpm-quota-cooldown-wait) own growth: open-sse/services/combo.ts 3548->3604 (+56, entirely this PR's diff — no other commit touched this file between the PR's merge-base and the release tip). Fixes combo cooldown-wait state recording so a bogus 503 is no longer crystallized when the cooldown-wait vars reset every setTry, adds an OpenAI-format SSE error frame path for combo-exhausted rejections (capturing request body + attempted models), and gives an abandoned per-target dispatch its own timeout instead of leaking a permanent 'pending' dashboard entry. Irreducible additions at the existing handleComboChat dispatch/retry chokepoint (mirrors the prior quota-share/headroom/task-aware strategy-branch precedents already frozen in this file). Covered by the PR's own combo-config + Gemini TPM-ceiling benchmark test additions.",
+    "_rebaseline_2026_07_22_8213_gemini_tpm_quota_cooldown_wait": "PR #8213 (hartmark, fix/gemini-tpm-quota-cooldown-wait) own growth: open-sse/services/accountFallback.ts 1857->1932 on the merged tip (release 1892 incl #8050 +35, plus this PR own growth +40); measured against the PR own merge-base was 1857->1898 (+41 — the release tip separately carries an unrelated +34 from #8050's antigravity 404 model-not-found lockout scoping, which this PR's branch does not include and this entry does not cover). Own growth is the Gemini TPM-ceiling classification + cooldown-wait wiring feeding into the combo cooldown-wait state machine (rate-limit wedge recovery) introduced by this PR's commit series. Irreducible additions at the existing account-fallback/model-lockout chokepoint. Covered by the PR's own gemini-rate-limit-tracker and TPM-ceiling benchmark test additions.",
+    "_rebaseline_2026_07_22_8213_health_unblock_model_cooldowns": "PR #8213 (hartmark, fix/gemini-tpm-quota-cooldown-wait) own growth: src/app/(dashboard)/dashboard/health/page.tsx 1094->1165 (+71, entirely this PR's diff — no other commit touched this file between the PR's merge-base and the release tip). Adds handleUnblockAll/handleUnblockOne dashboard actions (DELETE /api/resilience/model-cooldowns) so an operator can manually clear a Gemini TPM-wedge model lockout surfaced by this PR's cooldown-wait fixes, instead of waiting out the ceiling. Irreducible UI wiring at the existing health-page action chokepoint. Covered by the PR's own dashboard/resilience test additions.",
+    "_rebaseline_2026_07_22_8213_requestloggerdetail_unblock_ui": "PR #8213 (hartmark, fix/gemini-tpm-quota-cooldown-wait) own growth: src/shared/components/RequestLoggerDetail.tsx 799->941 (+142, entirely this PR's diff — no other commit touched this file between the PR's merge-base and the release tip; crosses the general 800-line new-file cap so is frozen here for the first time). Adds a collapsible section header (open/expand-less toggle) plus per-log-entry unblock (`unblocking`/`cleared` state, isCombo503 detection) so the request-logger detail panel surfaces the same Gemini TPM cooldown-wait / model-lockout unblock action introduced by this PR at the individual-request level (mirrors the health-page bulk unblock action added in the same PR). Covered by the PR's own dashboard/resilience test additions.",
+    "_rebaseline_2026_07_22_fusion_8013_8098_antigravity": "Fusion of #8013 (backryun, catalog/IDE-CLI-split rewrite) + #8098 (nguyenha935, protocol-fidelity/fail-closed/credits/tool-cloaking): open-sse/services/usage/antigravity.ts NEW 802 (>cap 800, +2 — #8098 credits/tier usage service on #8013's profile-aware headers). Test growth (models-catalog-route 1605->1608, provider-models-route 1752->1757 from #8013 Gemini 3.6 catalog) tracked in testFrozen.",
+    "_rebaseline_2026_07_23_8127_grok_weekly_quota": "#8127 (@apoapostolov) own growth: src/sse/handlers/chat.ts 1861->1865 (+4) — weekly quota tracking for grok-web wires a quota-fetch hook at the existing dispatch chokepoint. Thin wiring mirroring adjacent provider-quota branches; not extractable. Covered by tests/unit/grok-quota-fetcher.test.ts.",
+    "_rebaseline_2026_07_23_8143_empty_catch_logging": "#8143 (@chirag127) own growth: open-sse/utils/stream.ts 2869->2887 (+18) — replacing empty catch blocks in the SSE stream subsystem with console.debug logging (Rule #6 silent-swallow fix, issues #8138-#8142). Cohesive logging additions at the existing catch chokepoints, not extractable; frozen at new size. Covered by tests/unit/stream-handler-catch-logging-8143.test.ts.",
+    "_rebaseline_2026_07_23_8219_cache_ttl_settings_sidebar": "#8219 (@oyi77) own growth: sections.ts 806->813 (+7) — configurable model-catalog cache-TTL settings adds a new sidebar nav entry + its visibility wiring. Sidebar item array is data, not extractable logic; frozen at new size.",
+    "_rebaseline_2026_07_23_8247_8248_model_unhealthy": "#8247+#8248 own growth: accountFallback.ts 1940->1941 (+1, irreducible import statement only — the substantive #8248 DEGRADED-pattern classifier was extracted into open-sse/config/errorConfig.ts, which has ample headroom, instead of growing this frozen file; #8247's fix is a single existing-line condition change, net zero lines). Scoping the credits-exhausted 403/429 branch to isCompatibleProvider() (per-model-quota openai/anthropic-compatible-* nicknames) so it stays model-scoped instead of terminalling the whole connection, and classifying NVIDIA NIM 'Function ... DEGRADED' 400 bodies as model-access-denied instead of a raw passthrough 400. Covered by tests/unit/8247-accountfallback-model-unhealthy.test.ts and tests/unit/8248-accountfallback-nvidia-degraded.test.ts.",
+    "_rebaseline_2026_07_23_8252_combo_400_advance": "#8252 (@RaviTharuma) own growth: accountFallback.ts 1932->1940 (+8) + combo.ts 3604->3630 (+26) — advance combo on model-scoped 400s wrapped as invalid/Bad-Request. Irreducible wiring at existing account-fallback + combo dispatch chokepoints. Covered by combo-model-scoped-400-advance.test.ts.",
+    "_rebaseline_2026_07_23_8266_alibaba_media": "#8266 (@backryun) own growth: imageRegistry.ts 821->979 (+158) — Alibaba-family media models (Qwen image/video, Bailian, Wan) added to the image/video registry. Registry model data, not extractable logic; frozen at new size.",
+    "_rebaseline_2026_07_24_8388_compression_detail_persist": "#8388 (compression engine DETAIL settings — Headroom/session-dedup/CCR — dropped on save) own growth: src/lib/db/compression.ts 866->872 (+6 = irreducible call-site wiring at the existing getCompressionSettings/updateCompressionSettings chokepoint: one import line, one `...buildDetailConfigDefaults()` spread in the seed config, and one `case \"sessionDedup\": case \"ccr\": applyDetailConfigUpdate(config, key, parsed); break;` load-switch case, mirroring the existing headroom/#8056 case immediately above it). The actual normalizer logic (normalizeSessionDedupConfig/normalizeCcrConfig, matching SESSION_DEDUP_SCHEMA/CCR_SCHEMA bounds) was EXTRACTED into a new leaf src/lib/db/compressionDetailNormalizers.ts (well under cap) so this frozen file only carries the minimal dispatch wiring. Covered by tests/unit/8388-compression-detail-persist.test.ts (schema-accept + full DB save->reload round-trip for both new sub-objects, plus a no-regression assertion on the existing headroom round-trip).",
+    "_rebaseline_2026_07_24_responses_toolcalls_log_summary": "hartmark, fix/responses-tool-calls-log-summary own growth: open-sse/translator/response/openai-responses.ts 1163->1174 (+11). closeToolCall() now also writes the completed tool call into the shared state.toolCalls Map (already populated by the openai-to-claude / claude-to-openai / gemini-to-openai response translators) so stream.ts's completion-log summary builder (which reads state.toolCalls, not this translator's own funcCallIds/funcNames/funcArgsBuf bookkeeping) reports finish_reason \"tool_calls\" and message.tool_calls for openai->openai-responses translated streams instead of always logging \"stop\" with no tool_calls — the actual client-facing SSE events were already correct; only the persisted call-log summary was wrong. Irreducible call-site addition at the existing tool-call-close chokepoint. Covered by the new regression test in tests/unit/translator-resp-openai-responses.test.ts.",
+    "_rebaseline_2026_07_25_8476_combo_input_bound_homogeneous_scope": "PR #8476 (herjarsa, fix/8375-8459-combo-image-fixes, #8375) own growth: open-sse/services/combo.ts 3642->3679 (+37 net: +29 the PR's own isInputBoundFailure short-circuit for deterministic context_length_exceeded/context_window_exceeded failures, +8 a /green-prs pre-merge fix scoping that short-circuit to homogeneous remainders only — the shipped code fired unconditionally on ANY target, regressing the intentional heterogeneous-combo fallback #6637/isContextOverflow400 protects, exactly as flagged by this PR's own review evidence but never actually implemented in the branch). The fix compares orderedTargets[i+1..] modelStr against the failing target's modelStr at the existing executeTarget dispatch chokepoint (mirrors the sameProviderNext precedent a few lines below) — irreducible call-site wiring, not extractable without hiding the dispatch boundary. Covered by tests/unit/combo-input-bound-failure-8375.test.ts (homogeneous pool still short-circuits) and the new tests/unit/combo-input-bound-heterogeneous-8375.test.ts (heterogeneous combo now correctly falls through to the larger-context target).",
+    "_rebaseline_2026_07_25_adobe_firefly_reference_images": "Follow-up to #8006: storage upload + referenceBlobs for image/video and /v1/images/edits dispatch. adobeFireflyClient.ts 1958->2317 (+upload helpers, extract sources, resolve blob ids). Note: 2317 not 2316 — check-file-size.mjs counts LOC via split(\"\\n\").length (counts the trailing-newline empty element), which is 1 higher than `wc -l` on a file ending in \\n; the PR's original entry (2316) was measured with wc -l and undercounted by 1 against the actual gate.",
+    "_rebaseline_pr1043_minimax_tts": "Upstream port decolua/9router#1043 (toanalien) own growth: audioSpeech.ts 965->1061 (+96). Adds MiniMax T2A v2 TTS dispatch (handleMinimaxSpeech + hexToBytes helper) — provider entry was already in audioRegistry (format: minimax-tts) but no handler existed, falling through to the OpenAI-compatible default that fails (T2A has custom shape + hex-encoded audio + base_resp envelope). New branch sits next to the other inline provider branches (xiaomi-mimo, coqui, tortoise, aws-polly) — extracting would just create indirection. Covered by tests/unit/minimax-tts-1043.test.ts (3 tests, GREEN: success, base_resp error, invalid-hex).",
+    "_rebaseline_pr4592_exclude_exhausted_auto": "Reconcile #4592 already-merged growth: combo.ts 2991->3036 (+45, terminal-status quota-cutoff exclusion in buildAutoCandidates + opt-in gate). Fast-gate PR->release does not run check:file-size.",
+    "open-sse/executors/antigravity.ts": 1528,
+    "open-sse/executors/base.ts": 1640,
+    "open-sse/executors/chatgpt-web.ts": 3241,
+    "open-sse/executors/codex.ts": 1562,
+    "open-sse/executors/cursor.ts": 1560,
+    "open-sse/executors/deepseek-web.ts": 1148,
+    "open-sse/executors/grok-web.ts": 1044,
+    "open-sse/executors/muse-spark-web.ts": 1405,
+    "open-sse/handlers/chatCore.ts": 5034,
+    "open-sse/handlers/imageGeneration.ts": 3101,
+    "open-sse/handlers/responseSanitizer.ts": 1128,
+    "open-sse/handlers/search.ts": 1536,
+    "open-sse/handlers/videoGeneration.ts": 1063,
+    "open-sse/mcp-server/schemas/tools.ts": 1505,
+    "open-sse/mcp-server/server.ts": 1411,
+    "open-sse/mcp-server/tools/advancedTools.ts": 1120,
+    "open-sse/services/accountFallback.ts": 1972,
+    "open-sse/services/adobeFireflyClient.ts": 2385,
+    "open-sse/services/claudeCodeCompatible.ts": 1202,
+    "open-sse/services/combo.ts": 3648,
+    "open-sse/services/compression/strategySelector.ts": 1060,
+    "open-sse/services/rateLimitManager.ts": 1167,
+    "open-sse/translator/response/openai-responses.ts": 1204,
+    "open-sse/utils/cursorAgentProtobuf.ts": 1505,
+    "open-sse/utils/stream.ts": 2889,
+    "src/app/(dashboard)/dashboard/HomePageClient.tsx": 1388,
+    "src/app/(dashboard)/dashboard/analytics/ComboHealthTab.tsx": 1031,
+    "src/app/(dashboard)/dashboard/api-manager/ApiManagerPageClient.tsx": 3117,
+    "src/app/(dashboard)/dashboard/cache/media/MediaPageClient.tsx": 1067,
+    "src/app/(dashboard)/dashboard/combos/page.tsx": 4647,
+    "src/app/(dashboard)/dashboard/costs/CostOverviewTab.tsx": 1283,
+    "src/app/(dashboard)/dashboard/costs/quota-share/components/PoolWizard.tsx": 1022,
+    "src/app/(dashboard)/dashboard/endpoint/EndpointPageClient.tsx": 2615,
+    "src/app/(dashboard)/dashboard/health/page.tsx": 1165,
+    "src/app/(dashboard)/dashboard/providers/[id]/components/modals/EditConnectionModal.tsx": 1316,
+    "src/app/(dashboard)/dashboard/providers/page.tsx": 1944,
+    "src/app/(dashboard)/dashboard/runtime/RuntimePageClient.tsx": 1201,
+    "src/app/(dashboard)/dashboard/settings/components/PricingTab.tsx": 1019,
+    "src/app/(dashboard)/dashboard/settings/components/ProxyRegistryManager.tsx": 1464,
+    "src/app/(dashboard)/dashboard/settings/components/ResilienceTab.tsx": 1123,
+    "src/app/(dashboard)/dashboard/settings/components/RoutingTab.tsx": 1629,
+    "src/app/(dashboard)/dashboard/settings/components/SystemStorageTab.tsx": 1573,
+    "src/app/(dashboard)/dashboard/usage/components/BudgetTab.tsx": 1028,
+    "src/app/(dashboard)/dashboard/usage/components/EvalsTab.tsx": 2148,
+    "src/app/(dashboard)/dashboard/usage/components/ProviderLimits/index.tsx": 1119,
+    "src/app/api/providers/[id]/models/route.ts": 2250,
+    "src/app/api/v1/models/catalog.ts": 1549,
+    "src/lib/tokenHealthCheck.ts": 1021,
+    "src/lib/db/apiKeys.ts": 1529,
+    "src/lib/db/core.ts": 1637,
+    "src/lib/db/migrationRunner.ts": 1084,
+    "src/lib/db/models.ts": 1097,
+    "src/lib/db/providers.ts": 1034,
+    "src/lib/memory/retrieval.ts": 1073,
+    "src/lib/tailscaleTunnel.ts": 1202,
+    "src/lib/usage/providerLimits.ts": 1013,
+    "src/shared/components/OAuthModal.tsx": 1134,
+    "src/shared/components/RequestLoggerV2.tsx": 1629,
+    "src/shared/components/analytics/charts.tsx": 1035,
+    "src/shared/services/cliRuntime.ts": 1122,
+    "src/sse/handlers/chat.ts": 1877,
+    "src/sse/services/auth.ts": 2508,
+    "tests/unit/account-fallback-service.test.ts": 1572,
+    "tests/unit/provider-validation-specialty.test.ts": 2985,
+    "open-sse/executors/hyperagent.ts": 1026,
+    "open-sse/executors/default.ts": 1042,
+    "open-sse/executors/kiro.ts": 1069
+  },
   "_rebaseline_2026_07_27_v3849_train2": "Merge-train 2 (7 PRs) — owner-approved 2026-07-27. Single entry: chatCore.ts 4955->5006 (#8595, Responses multi-turn image compaction before the context hard-reject). Genuine irreducible growth at the existing compaction chokepoint in handleChatCore — the PR adds a last-resort retry against the concrete budget plus the estimateFinalInputTokens helper, both wired at the pre-existing call site rather than a new branch. Covered by tests/unit/8560-responses-image-compaction.test.ts (4 tests).",
   "_rebaseline_2026_07_27_v3849_train3": "Merge-train 3 (13 PRs) — owner-approved 2026-07-27. Both entries are genuine irreducible growth at existing chokepoints, not new branches: src/lib/db/apiKeys.ts 1518->1529 (#8805 cx/* ≡ codex/* API-key model permissions); open-sse/handlers/chatCore.ts 5006->5020 (#8806 real response payload into plugin onResponse hooks). Covered by tests/unit/db-apiKeys-crud.test.ts (4 new cases) and the two plugin-hook test files updated in #8806 respectively.",
   "_rebaseline_2026_07_28_8842_antigravity_projectid_refresh": "PR #8842 (fix/antigravity-projectid-refresh) own growth: open-sse/executors/antigravity.ts 1493->1528 (+35 = projectId discovery in refreshCredentials: import ensureAntigravityProjectAssigned + trim projectId + call ensureAntigravityProjectAssigned with 8s timeout + persistDiscoveredAntigravityProjectId + log success/failure). Irreducible wiring at the existing credential-refresh chokepoint. Covered by tests/unit/executor-antigravity.test.ts (4 new test cases).",

config/quality/open-sse-typecheck-baseline.json

@@ -1,176 +0,0 @@
-{
-  "open-sse/executors/azure-openai.ts": {
-    "TS2345": 1
-  },
-  "open-sse/executors/chatgpt-web.ts": {
-    "TS2339": 1
-  },
-  "open-sse/executors/claude-web/stream.ts": {
-    "TS2322": 1,
-    "TS2345": 1
-  },
-  "open-sse/executors/copilot-web.ts": {
-    "TS2353": 1
-  },
-  "open-sse/executors/deepseek-web.ts": {
-    "TS2352": 1
-  },
-  "open-sse/executors/default.ts": {
-    "TS2352": 1
-  },
-  "open-sse/executors/duckduckgo-web.ts": {
-    "TS2345": 2
-  },
-  "open-sse/executors/duckduckgo-web/challenge.ts": {
-    "TS2304": 1
-  },
-  "open-sse/executors/edgeTts.ts": {
-    "TS2345": 1
-  },
-  "open-sse/executors/gemini-business.ts": {
-    "TS2339": 1
-  },
-  "open-sse/executors/ghe-copilot.ts": {
-    "TS2554": 1
-  },
-  "open-sse/executors/inner-ai.ts": {
-    "TS2352": 2
-  },
-  "open-sse/executors/theoldllm.ts": {
-    "TS2322": 1
-  },
-  "open-sse/executors/veoaifree-web.ts": {
-    "TS2322": 1
-  },
-  "open-sse/executors/windsurf.ts": {
-    "TS2322": 1
-  },
-  "open-sse/handlers/chatCore.ts": {
-    "TS2339": 30,
-    "TS2322": 1,
-    "TS2345": 11
-  },
-  "open-sse/handlers/chatCore/claudeUpstreamMessages.ts": {
-    "TS2345": 1
-  },
-  "open-sse/handlers/chatCore/clientUsageBuffer.ts": {
-    "TS2345": 1
-  },
-  "open-sse/handlers/chatCore/clineResponseEnvelope.ts": {
-    "TS2698": 1
-  },
-  "open-sse/handlers/chatCore/compressionAnalyticsWrite.ts": {
-    "TS2724": 1
-  },
-  "open-sse/handlers/chatCore/nonStreamingResponseHeaders.ts": {
-    "TS2322": 2
-  },
-  "open-sse/handlers/chatCore/sanitization.ts": {
-    "TS2339": 1,
-    "TS2537": 1
-  },
-  "open-sse/handlers/chatCore/semanticCacheStore.ts": {
-    "TS2345": 1
-  },
-  "open-sse/handlers/chatCore/streamingPipeline.ts": {
-    "TS2345": 2
-  },
-  "open-sse/handlers/chatCore/streamingSemanticCacheStore.ts": {
-    "TS2345": 1
-  },
-  "open-sse/handlers/chatCore/thinkingSignatureRecovery.ts": {
-    "TS2339": 2
-  },
-  "open-sse/handlers/imageGeneration.ts": {
-    "TS2554": 2
-  },
-  "open-sse/handlers/responsesHandler.ts": {
-    "TS2339": 1,
-    "TS2345": 1
-  },
-  "open-sse/handlers/sseParser.ts": {
-    "TS2322": 2
-  },
-  "open-sse/handlers/videoGeneration.ts": {
-    "TS2339": 2
-  },
-  "open-sse/mcp-server/tools/compressionTools.ts": {
-    "TS2339": 2
-  },
-  "open-sse/services/__tests__/specificityDetector.test.ts": {
-    "TS2353": 2
-  },
-  "open-sse/services/browserBackedChat.ts": {
-    "TS2322": 1,
-    "TS2794": 1
-  },
-  "open-sse/services/claudeAdaptiveThinking.ts": {
-    "TS2352": 2
-  },
-  "open-sse/services/comboManifestMetrics.ts": {
-    "TS2307": 1
-  },
-  "open-sse/services/compression/engines/ccr/index.ts": {
-    "TS2339": 1
-  },
-  "open-sse/services/payloadRules.ts": {
-    "TS2677": 1
-  },
-  "open-sse/services/tokenLimitCounter.ts": {
-    "TS2551": 1
-  },
-  "open-sse/transformer/responsesTransformer.ts": {
-    "TS2339": 1
-  },
-  "open-sse/utils/stream.ts": {
-    "TS2339": 7,
-    "TS2345": 1,
-    "TS2556": 1
-  },
-  "src/app/api/v1/_shared/mediaGenerationRoute.ts": {
-    "TS2339": 2
-  },
-  "src/app/api/v1/models/catalog.ts": {
-    "TS2345": 1
-  },
-  "src/app/api/v1/models/catalogVision.ts": {
-    "TS2322": 1
-  },
-  "src/app/api/v1/videos/generations/route.ts": {
-    "TS2322": 1,
-    "TS2345": 1
-  },
-  "src/lib/guardrails/visionBridge.ts": {
-    "TS2345": 1
-  },
-  "src/lib/providers/codexFastTier.ts": {
-    "TS2367": 1
-  },
-  "src/lib/skills/builtins.ts": {
-    "TS2322": 1
-  },
-  "src/lib/skills/injection.ts": {
-    "TS2339": 1
-  },
-  "src/lib/skills/webFetchExecution.ts": {
-    "TS2322": 1
-  },
-  "src/lib/streamingPiiTransform.ts": {
-    "TS2345": 1
-  },
-  "src/shared/providers/webSessionCredentials.ts": {
-    "TS2353": 1,
-    "TS2322": 1
-  },
-  "src/shared/validation/helpers.ts": {
-    "TS2339": 1
-  },
-  "src/sse/handlers/chat.ts": {
-    "TS2352": 1,
-    "TS2322": 2,
-    "TS2339": 1
-  },
-  "src/sse/services/model.ts": {
-    "TS2339": 4
-  }
-}

config/quality/quality-baseline.json

@@ -110,8 +110,9 @@
       "_rebaseline_2026_06_26_v3837_release": "343->345. v3.8.37 cycle drift surfaced by the release-green pre-flight (the Quality Ratchet does NOT run on PR->release fast-gates, so warnings/complexity accrued unmeasured across this cycle's 76 commits — provider adds DGrid/Pioneer/xAI, headroom proxy lifecycle #4649, ~50 SSE/translator fixes, Engine Combos #5062). Trust-but-verify: this release-finalize working tree touches ONLY CHANGELOG.md, docs/i18n/*/CHANGELOG.md mirrors, and these baselines — 0 production-code change, so all drift is inherited cycle drift (`any` warn-allowed in open-sse/ + tests/). Tighten via --require-tighten next cycle."
     },
     "cognitiveComplexity": {
+      "value": 957,
+      "_rebaseline_2026_07_25_dario_upstream_proxy_selector": "951->957 (+6). Same cycle-drift + own-growth split as the complexity-baseline.json note dated 2026-07-25 (PR #8523, Dario embedded service): cognitive-complexity does not run on PR->release fast-gates, so drift accrues unratcheted. Base upstream/release/v3.8.49 tip measures 956 locally with this PR\u0027s commits removed; this branch measures 957 both locally and on the CI runner. This PR\u0027s own genuine contribution is +1: the new mode-selector conditional rendering (Native/CLIProxyAPI/Dario/Fallback branches plus the fallback-backend picker) in ConnectionRow.tsx. Structural shrink stays tracked in #3501. Tighten via --update next cycle.",
       "value": 1223,
-      "_rebaseline_2026_07_25_dario_upstream_proxy_selector": "951->957 (+6). Same cycle-drift + own-growth split as the complexity-baseline.json note dated 2026-07-25 (PR #8523, Dario embedded service): cognitive-complexity does not run on PR->release fast-gates, so drift accrues unratcheted. Base upstream/release/v3.8.49 tip measures 956 locally with this PR's commits removed; this branch measures 957 both locally and on the CI runner. This PR's own genuine contribution is +1: the new mode-selector conditional rendering (Native/CLIProxyAPI/Dario/Fallback branches plus the fallback-backend picker) in ConnectionRow.tsx. Structural shrink stays tracked in #3501. Tighten via --update next cycle.",
       "_rebaseline_2026_07_25_8470_hyperagent_sticky_thread": "951->957 (+6). PR #8470 (artickc, fix/hyperagent-tool-loop-thread-sticky) pre-green validation. Trust-but-verify: origin/release/v3.8.49 tip alone (pristine, no PR changes) already measures 956 with node scripts/check/check-cognitive-complexity.mjs — i.e. +5 is inherited cycle drift unrelated to this PR (cognitive-complexity does not run on PR->release fast-gates). This PR's OWN growth adds exactly +1: per-file eslint scoped scan (eslint --config eslint.complexity-ratchets.config.mjs open-sse/executors/hyperagent.ts) on base vs PR shows extractMessageText() crossing the threshold for the first time (new sonarjs/cognitive-complexity violation, 26 > 15) from the new Anthropic tool_use/tool_result flattening branches; resolveHyperAgentThreadBinding's existing pre-#8470 violation (16) grows to 21 (still counted once) from the new root-key lookup tier; createHyperAgentThread and execute() are unchanged pre-existing violations. Net repo-wide total = 956 (inherited drift) + 1 (this PR's own new violation) = 957. Full-repo re-measurement of the merged branch was attempted but not completed live due to heavy concurrent devbox load (many other /green-prs sessions running the identical full-repo eslint scan in parallel); the value here is derived from two independently-clean measurements (base-tip full scan + per-file base-vs-PR delta) rather than a third full-repo run. Covered by tests/unit/executor-hyperagent.test.ts (19/19). Tighten via --update next cycle.",
       "_rebaseline_2026_07_25b_v3849_mergetrain_owngrowth": "Owner-approved (chat, 2026-07-25): 956->968 (+12). v3.8.49 /merge-prs 41-PR merge-train aggregate own-growth: measured 968 on the combined boarded tree (tip ac15014ca7) vs 956 on the pristine release tip. The batch's new over-threshold functions come from the pre-screen-flagged complexity-growth set (#8378/#8432/#8476/#8526 etc); each PR is under-ceiling alone, the combined batch adds +12. Same merge-burst class as the notes below; owner chose ceiling-absorb over per-PR extraction. Structural shrink tracked in #3501; tighten via --update next cycle.",
       "_rebaseline_2026_07_25_v3849_mergequeue_drain": "Owner-approved (chat, 2026-07-25): 951->956 (+5). v3.8.49 /merge-prs queue-drain: inherited cognitive-complexity drift from the cycle's merge burst (base-red slices + owner PRs + parallel-session merges #8500-8508); check:cognitive-complexity does not run on PR->release fast-gates, so it accrued unmeasured. Measured 956 on the pristine release tip 4053e2314a alone (BEFORE any queue PR boards) — the entire +5 is base drift already on the tip, reddening Fast Quality Gates for every merge-ready PR. Owner approved raising the ceiling to the measured tip value so the ~34-PR merge-train lands without per-PR extraction churn. Structural shrink tracked in #3501; tighten via --update next cycle.",
@@ -147,10 +148,9 @@
       "dedicatedGate": true
     },
     "codeqlAlerts": {
-      "value": 1,
+      "value": 0,
       "direction": "down",
-      "dedicatedGate": true,
-      "_rebaseline_2026_08_06_base_grew": "Base branch file-size drift: translator-openai-to-gemini.test.ts grew 1619->1622 (test assertions for Gemini translator compatibility). CodeQL alert (js/insufficient-password-hash in raycast.ts) is pre-existing base-red; incremented baseline to match."
+      "dedicatedGate": true
     },
     "secretFindings": {
       "_note": "Zeroed 2026-07-13 (WS6/D3): the 3 frozen generic-api-key FPs are allowlisted with justification in .gitleaks.toml — any NEW finding regresses the ratchet.",

docs/compression/COMPRESSION_ENGINES.md

@@ -183,11 +183,6 @@ Per environment:
   ships slim by design.
 - **VPS (PM2)** — install into the app's `node_modules`, then restart the process so the
   worker re-probes the gate.
-- **Raw Next standalone (`npm run build` → `.build/next/standalone/server.js`)** — the
-  standalone trace ships NEITHER the worker nor the optional deps, so the engine silently
-  fail-opens. `scripts/build/colocate-standalone.mjs` re-applies both (worker esbuild +
-  optional-dep closure into the standalone tree); it runs automatically via the
-  `postbuild` npm hook after every build. Idempotent, fail-soft when deps are absent.
 
 **Verify it is active:** with LLMLingua selected, real prose actually shrinks (the engine
 stops fail-opening), and the first request triggers the model download into

docs/compression/COMPRESSION_GUIDE.md

@@ -146,26 +146,6 @@ That `78-95%` number applies when both RTK and Caveman can reduce the same input
 Caveman response output mode is separate: when enabled, use Caveman's own output savings (`65%`
 average, `~75%` headline, `22-87%` range). Total billing savings depend on your prompt/output mix.
 
-### What "eligible" actually means
-
-The 15-95% headline range is real, but it only applies to **redundant or verbose** content — repeated
-error lines, a build log that spams the same warning, an oversized `grep`/file-read dump. It does
-**not** mean every request saves that much.
-
-Verified empirically (`tests/unit/compression/stacked-compression-tool-result-savings.test.ts`): a
-`stacked` (RTK + Caveman) run against an Anthropic-shape `tool_result` block containing 300 identical
-error lines produced **95.93% token savings / 96.26% character savings** — squarely in the advertised
-range. But the same pipeline run against normal, non-redundant tool output (a clean `grep` match list,
-a short file read, ordinary conversational text) correctly produces **near-zero savings**, because
-there is nothing repetitive to remove and `validateCompression()` (`validation.ts`) refuses to ship a
-rewrite that would drop or alter code blocks, URLs, headings, versions, or `CONST_CASE` identifiers.
-
-This is expected, safe behavior, not a bug: a coding session that mostly reads/greps clean files will
-see modest total savings even with compression fully enabled, while a session that hits a failing
-loop or a chatty linter will see the full 78-95% range on that traffic. Don't use a single session's
-low aggregate savings percentage as evidence compression is misconfigured — check whether the
-underlying tool output was actually redundant first.
-
 ---
 
 ## Token Savings Visualization

docs/frameworks/RADAR.md

@@ -1,17 +1,17 @@
 ---
 title: "Radar Free-Model Catalog"
 version: 3.8.50
-lastUpdated: 2026-08-07
+lastUpdated: 2026-08-05
 ---
 
 # Radar Free-Model Catalog
 
 > **Source of truth:** `src/lib/radar/`, `src/lib/db/radar.ts`, `src/app/api/radar/`
-> **Last updated:** 2026-08-07 — v3.8.50
+> **Last updated:** 2026-08-05 — v3.8.50
 
 Radar is an **optional add-on** that overlays a signed, freshly-curated free-model
 catalog on top of the release baseline (`FREE_MODEL_BUDGETS` in
-`open-sse/config/freeModelCatalog.data.ts`). It exists because the free-tier landscape moves
+`open-sse/config/freeModelCatalog.ts`). It exists because the free-tier landscape moves
 faster than release cadence — providers add, shrink, or discontinue free quotas between
 releases, and the baseline catalog can only be refreshed when a new version ships.
 
@@ -82,43 +82,6 @@ that lets the feed service decide which tier to serve (see
 
 ---
 
-## Getting a supporter key
-
-The activation screen (`/dashboard/radar`) links out to two flows for **obtaining** a
-supporter key. The OSS repo itself never issues one, never runs payment code, and
-**never states a price** — pricing is decided and displayed entirely on the
-destination pages, not in this repo (spec decision D14).
-
-- **"I'm a contributor"** — opens `RADAR_CONTRIBUTOR_CLAIM_URL` (default
-  `https://radar.omniroute.online/auth/github`), a GitHub OAuth claim flow hosted on
-  the private radar server. It verifies the visitor's GitHub account and grants a
-  supporter key to anyone with 5+ merged pull requests or a top-100 contributor spot
-  on the repo.
-- **"Support the project"** — opens `RADAR_SUPPORTER_PLANS_URL` (default
-  `https://radar.omniroute.online/planos`), the payment/plans page.
-
-Both URLs are resolved server-side (`src/lib/radar/links.ts`, same env-override
-pattern as `RADAR_FEED_URL`) and relayed to the dashboard through the existing
-`GET /api/radar/settings` response (`contributorClaimUrl`, `supporterPlansUrl`) — the
-client component never reads `process.env` itself.
-
-| Var                             | Purpose                                                                                      |
-| -------------------------------- | ---------------------------------------------------------------------------------------------- |
-| `RADAR_CONTRIBUTOR_CLAIM_URL`   | Overrides the contributor-claim URL (default `https://radar.omniroute.online/auth/github`).   |
-| `RADAR_SUPPORTER_PLANS_URL`     | Overrides the supporter-plans URL (default `https://radar.omniroute.online/planos`).          |
-
-Once a visitor has a key (`omr_` + 40 hex chars), it is set with `POST
-/api/radar/settings` (`{ supporterKey }`) — the same endpoint documented under
-[Data sync](#data-sync-is-a-separate-opt-in--the-privacy-promise) above.
-
-**Known gap:** the dashboard activation screen does not yet have a dedicated
-key-paste input — pasting a key today requires calling `POST /api/radar/settings`
-directly (curl, a script, or a future UI). This release only adds the two claim/plans
-buttons; the API already accepts and masks the key, but no `<input>` for it exists in
-`src/app/(dashboard)/dashboard/radar/page.tsx` yet.
-
----
-
 ## Security model
 
 ### Ed25519 signature over exact bytes
@@ -164,24 +127,6 @@ untouched. The cached payload is defensively re-validated again on every read
 (`getRadarCatalog()`) — a corrupted or hand-edited cache row falls back to the
 baseline rather than being served.
 
-### Response size cap (10 MB)
-
-`syncRadar()` enforces a **10 MB hard cap** on the feed response body — the signed
-feed is a KB-scale JSON document, so anything past this points at a misconfigured or
-hostile `RADAR_FEED_URL` (or an upstream serving garbage), not a legitimate catalog.
-Enforcement is two-layered:
-
-1. A `Content-Length` preflight check skips reading the body entirely when the
-   header already declares a value over the cap.
-2. A running-total check while reading the body enforces the cap even when
-   `Content-Length` is absent or understates the real size — the header is never
-   trusted on its own. Concatenating the accumulated chunks preserves the exact
-   bytes needed for the Ed25519 signature check afterward.
-
-Exceeding the cap returns `{ status: "too_large" }` and leaves the cache untouched,
-following the same non-destructive pattern as every other sync failure
-(`invalid_signature`, `invalid_schema`, `stale`).
-
 ---
 
 ## Tiers: `community` and `live`
@@ -201,28 +146,6 @@ recoverable, all non-fatal to the cached state) from a successful `{ status:
 "updated", version, tier }`. There is no tier-specific error path a client needs to
 handle.
 
-### The served tier comes from a response header, not the signed body
-
-The signed feed **body**'s `tier` field is always `"live"` — the feed service ships
-**one signed artifact per version**, so the body cannot carry a per-request tier
-without invalidating the Ed25519 signature (re-signing per request would defeat the
-point of a pinned, cacheable, verifiable artifact). The tier actually served for a
-given request is instead carried in the **`x-omniroute-feed-tier` response header**,
-decided server-side from the request's `Authorization` key.
-
-`syncRadar()` (`src/lib/radar/sync.ts::parseServedTierHeader()`) is the single place
-that resolves the tier a client should trust:
-
-1. Parse `x-omniroute-feed-tier` with `RadarTierSchema` (Zod) — an absent header, or
-   a value that isn't exactly `"community"` or `"live"`, is treated as **not
-   present** (never trusted into the cache/UI as-is; this also covers older feed
-   servers that predate the header).
-2. Fall back to the signed body's `tier` field (always `"live"`) only when step 1
-   yields nothing.
-3. The resolved tier is what gets cached and returned as `{ status: "updated",
-   version, tier }` — this is the value the dashboard shows, never the raw body
-   field.
-
 ---
 
 ## Read-time overlay merge rules
@@ -260,15 +183,13 @@ Every merged entry carries an `origin` field the UI renders as a badge:
 
 ## Local surfaces — never a feed proxy
 
-Five local routes back the UI, all under `src/app/api/radar/`:
+Three local routes back the UI, all under `src/app/api/radar/`:
 
-| Route                   | Method | Purpose                                                                                          |
-| ----------------------- | ------ | -------------------------------------------------------------------------------------------------- |
-| `/api/radar/catalog`    | GET    | Returns the merged catalog (`getRadarCatalog()`) from the local cache.                            |
-| `/api/radar/sync`       | POST   | Triggers `syncRadar()` server-side; returns the resulting status.                                 |
-| `/api/radar/settings`   | GET    | Returns `{ optIn, hasSupporterKey, supporterKeyMasked }` — never the raw key.                     |
-| `/api/radar/settings`   | POST   | Sets opt-in and/or the (encrypted) supporter key.                                                 |
-| `/api/radar/referrals`  | GET    | Returns `{ fixed, campaigns, tier }` from the local cache — see [Referral links](#referral-links-free-credits) below. |
+| Route                 | Method | Purpose                                                                |
+| --------------------- | ------ | ---------------------------------------------------------------------- |
+| `/api/radar/catalog`  | GET    | Returns the merged catalog (`getRadarCatalog()`) from the local cache. |
+| `/api/radar/sync`     | POST   | Triggers `syncRadar()` server-side; returns the resulting status.      |
+| `/api/radar/settings` | POST   | Sets opt-in and/or the (encrypted) supporter key.                      |
 
 **Hard rule: these routes never proxy the feed service.** The browser only ever talks
 to the local OmniRoute server; `syncRadar()` is the single module in the whole client
@@ -276,118 +197,11 @@ that touches the network for Radar (`src/lib/radar/sync.ts`), and it always runs
 server-side, never client-side. This keeps the feed URL and any supporter key
 out of client-facing network traffic entirely.
 
-All five routes return `404` when `RADAR_ENABLED` is off (see
+All three routes return `404` when `RADAR_ENABLED` is off (see
 [Flag](#flag-radar_enabled-default-off) above), and route error responses through
 `buildErrorBody()`/`sanitizeErrorMessage()` per the repo-wide error-sanitization rule
 (`docs/security/ERROR_SANITIZATION.md`).
 
-### Authentication
-
-All five routes require authentication via `isAuthenticated()`
-(`src/shared/utils/apiAuth.ts`) — a dashboard session cookie or a management-scoped
-API key, the same gate that protects the rest of `/api/settings/*`. The flag-off
-`404` check always runs **before** the auth check, so an install with `RADAR_ENABLED`
-off stays byte-identical (no auth prompt just to learn the surface doesn't exist);
-once the flag is on, an unauthenticated request gets `401` before any DB read or
-write. `GET /api/radar/settings` never returns the raw supporter key regardless of
-auth state — only the masked form and a `hasSupporterKey` boolean.
-
----
-
-## Referral links (free credits)
-
-The server-published feed carries a `referrals` section (server-side D28 work, already
-in production — this section documents the **client** consumption only):
-
-```ts
-referrals: {
-  fixed: RadarReferral[],      // present in EVERY tier, including community
-  campaigns: RadarReferral[],  // only populated on the live (supporter) tier;
-                                // the community artifact always publishes []
-}
-// RadarReferral = { provider, url, kind: "fixo" | "campanha", validUntil,
-//                    requiredAction, isDefault }
-```
-
-The client never decides which tier it received or which referrals belong in which
-tier — the server already publishes two artifacts (`live`/`community`) with
-`campaigns` gated server-side, same principle as the [tiers](#tiers-community-and-live)
-section above. `RadarFeedSchema` (`src/lib/radar/feedSchema.ts`) validates `referrals`
-as a whole-object `.default({fixed:[],campaigns:[]})`, and `campaigns` defaults
-independently inside it — so a feed cached before this section existed on the server
-still parses cleanly, and `campaigns` alone can also be absent without failing
-validation. Every `RadarReferral.url` must be `https://` — a `http://` url fails
-schema validation.
-
-### Accessor
-
-`src/lib/radar/index.ts` exports two read-only accessors, both never throwing (same
-defensive contract as `getRadarCatalog()` — flag off, no cache, or a corrupt/old cached
-payload all resolve to the empty shape instead of an error):
-
-- `getRadarReferrals()` → `{ fixed: RadarReferral[], campaigns: RadarReferral[] }`.
-- `getDefaultReferralFor(provider)` → the `fixed` referral with `isDefault: true` for
-  that provider, or `null`. Only looks at `fixed` — a campaign is never used as a
-  provider's "default" link.
-
-The actual "which referral is the default for a provider" rule lives in
-`findDefaultReferral()` (`src/lib/radar/referrals.ts`), a small pure function with **no
-DB import** — it is safe to import into a `"use client"` component. `getRadarReferrals`/
-`getDefaultReferralFor` (in `index.ts`) pull in `@/lib/db/radar` and therefore stay
-server-only; the providers dashboard imports `referrals.ts` directly instead of
-`index.ts` (see below) to avoid bundling `better-sqlite3` into the browser.
-
-### `GET /api/radar/referrals`
-
-Follows the exact same gate order as every other Radar route: `RADAR_ENABLED` off →
-`404` (checked first, byte-identical inertia); unauthenticated → `401`; otherwise `200`
-with `{ fixed, campaigns, tier }` — `tier` comes straight from the cache row and is
-purely informative (drives the UI's soft upsell copy below), the route does no
-gating of its own. Never proxies the feed server — same local-cache-only contract as
-`/api/radar/catalog`.
-
-### Dashboard UI — "Free credits" tab on `/dashboard/radar`
-
-Reuses the existing Radar page (`src/app/(dashboard)/dashboard/radar/page.tsx`) as a
-second tab instead of a new route — less routing/i18n surface for a feature that is a
-variation on data the page already fetches. Once opted in, the tab bar offers
-**Catalog** (existing table) and **Free credits**:
-
-- Fixed links are grouped by provider, each showing `requiredAction` (when present)
-  and a `target="_blank" rel="noopener noreferrer"` button to the referral URL.
-- Campaigns show the same, plus `validUntil` when present.
-- When `campaigns` is empty **and** the served tier is `community`, the UI shows a
-  short upsell note ("limited-time campaigns are a supporter extra") — this **never**
-  hides or gates the fixed links list, which stays fully populated for every tier. The
-  upsell is soft messaging only, never a block.
-
-### Referral link on the provider name (providers dashboard)
-
-`ProviderPageHeader` (`src/app/(dashboard)/dashboard/providers/[id]/components/`)
-already linked the provider name to `providerInfo.website` when present, with one
-precedent for a monetized link: the Kimi (Moonshot AI) partner-link note
-(`providers.kimiPartnerLinkNote` i18n key). D28 reuses that exact same discreet-note
-pattern for Radar default referrals instead of introducing a new key.
-
-Loose coupling, by design:
-
-- `resolveProviderHeaderLink()` (`src/app/(dashboard)/dashboard/providers/providerPageUtils.ts`)
-  is a **pure** function — `(staticWebsite, referralUrl) => { website, isReferralLink }`
-  — with no dependency on `@/lib/radar` or `@/lib/db/*`. `providerPageUtils.ts` as a
-  whole stays free of those imports (asserted by
-  `tests/unit/provider-header-referral-link.test.ts`).
-- `ProviderDetailPageClient.tsx` (a `"use client"` component) is the one place allowed
-  to fetch Radar data — via `fetch("/api/radar/referrals")`, the same local-route
-  pattern the Radar dashboard page itself uses — and it computes the default referral
-  client-side with `findDefaultReferral()` from the DB-free `src/lib/radar/referrals.ts`.
-- With `RADAR_ENABLED` off, the fetch 404s, `referralUrl` stays `null`, and
-  `resolveProviderHeaderLink()` returns the static catalog `website` unchanged — the
-  provider page is byte-identical to before this feature existed. Same outcome when
-  there is no cache yet or no default referral for that specific provider.
-- When a default referral does apply, `ProviderPageHeader` receives `isReferralLink`
-  and shows the same discreet note/tooltip as the Kimi partner link (reusing the
-  `providers.kimiPartnerLinkNote` key) — never a new, separate visual treatment.
-
 ---
 
 ## How to self-host a feed
@@ -417,6 +231,6 @@ feed.
 ## Related docs
 
 - [`docs/security/ERROR_SANITIZATION.md`](../security/ERROR_SANITIZATION.md) — the
-  error-response pattern the five `/api/radar/*` routes follow.
+  error-response pattern the three `/api/radar/*` routes follow.
 - [`docs/reference/ENVIRONMENT.md`](../reference/ENVIRONMENT.md#27-radar-feed-self-hosting)
   — `RADAR_FEED_URL` / `RADAR_FEED_PUBKEY` reference.

docs/guides/TROUBLESHOOTING.md

@@ -489,59 +489,6 @@ Provider profiles support these settings:
 
 When many concurrent requests hit a rate-limited provider, OmniRoute uses mutex + auto rate-limiting to serialize requests and prevent cascading failures. This is automatic for API key providers.
 
-### Chat requests fail with 503 / chat_admission_busy
-
-**Symptoms:**
-
-- The chat completions endpoint returns a retryable `503` response whose error code is
-  `chat_admission_busy`.
-- The response includes `Retry-After`; the byte-based path uses 2 seconds, while the
-  structure-based path uses 1 second and includes `reason: "structure_limit"`.
-- This can happen while another heavyweight chat or long-running streaming response is still
-  in flight.
-
-The byte-based response body is:
-
-```json
-{
-  "error": {
-    "message": "Chat admission capacity is temporarily unavailable. Retry shortly.",
-    "type": "server_error",
-    "code": "chat_admission_busy"
-  }
-}
-```
-
-The structure-based response uses the same type and code, with the message
-`Structurally heavy chat request capacity is busy; retry shortly.` and
-`reason: "structure_limit"`.
-At the default thresholds, a request is structurally heavy when it has at least `200` messages,
-at least `64` tools, or at least `32,000` estimated tokens, or when bounded structure estimation
-exhausts its bounds of `10,000` visited nodes or depth `12`.
-
-**Cause:** This is deliberate load shedding inside OmniRoute, not an upstream-provider failure.
-Each process uses a process-local guard to reserve limited heavyweight capacity before retaining
-and parsing a large request body. A heavyweight lease remains held for the lifetime of an SSE
-response.
-Current heavyweight lease occupancy is not surfaced in the dashboard.
-Settings → Resilience → Request Queue → Concurrent Requests does not control this; that setting
-governs a separate provider request-queue mechanism.
-
-**Fix:**
-
-1. Retry first. Clients should honor `Retry-After` and use backoff rather than immediately
-   repeating the request.
-2. If normal deployment traffic repeatedly exhausts the guard, you can cautiously raise
-   `OMNIROUTE_CHAT_MAX_HEAVY_IN_FLIGHT` from its default of `1`. Increase it one step at a time,
-   restart OmniRoute after each change, and observe memory headroom under representative load.
-   Every additional heavyweight request can increase concurrent V8 heap use and container or
-   host OOM risk. No value is safe for every deployment; validate the setting against your own
-   traffic and memory limits rather than assuming that `2` is universally safe.
-
-See the [environment-variable reference](../reference/ENVIRONMENT.md#4-security--authentication)
-for the authoritative admission settings. Loosening the heavyweight classification thresholds
-can let expensive requests bypass this guard and is riskier than a cautious in-flight increase.
-
 ---
 
 ## Optional RAG / LLM failure taxonomy (16 problems)

docs/reference/ENVIRONMENT.md

@@ -194,7 +194,7 @@ OmniRoute uses **SQLite** (via `better-sqlite3`) for all persistence. These vari
 | `OMNIROUTE_CHAT_HEAVY_MESSAGE_COUNT`    | `200`                   | `src/shared/middleware/chatBodyAdmission.ts`     | Message count that classifies a chat request as heavyweight even when its body is below the byte threshold.                                                                                                                                                                                                                                                        |
 | `OMNIROUTE_CHAT_HEAVY_TOOL_COUNT`       | `64`                    | `src/shared/middleware/chatBodyAdmission.ts`     | Tool count that classifies a chat request as heavyweight even when its body is below the byte threshold.                                                                                                                                                                                                                                                           |
 | `OMNIROUTE_CHAT_HEAVY_ESTIMATED_TOKENS` | `32000`                 | `src/shared/middleware/chatBodyAdmission.ts`     | Conservative string-size token estimate that classifies a request as heavyweight; this is an admission-cost proxy, not provider billing tokenization.                                                                                                                                                                                                              |
-| `OMNIROUTE_CHAT_HARD_MAX_MESSAGES`      | `0` (disabled)          | `src/shared/middleware/chatBodyAdmission.ts`     | Optional opt-in chat history cap. Disabled by default: a message count is deployment policy, not a universal property of a request, and capping here rejects conversations with a terminal `413` before the compression pipeline can make them servable. Heap growth is bounded by `OMNIROUTE_CHAT_MAX_HEAVY_IN_FLIGHT` and the heap-pressure shed. Set a positive value on memory-constrained deployments that need a hard ceiling; excess then receives structured compact-required `413`. |
+| `OMNIROUTE_CHAT_HARD_MAX_MESSAGES`      | `800`                   | `src/shared/middleware/chatBodyAdmission.ts`     | Hard chat history cap. Requests above it receive structured compact-required `413` before compression, translation, or provider dispatch.                                                                                                                                                                                                                           |
 | `OMNIROUTE_MAX_NONSTREAMING_RESPONSE_BYTES` | `67108864` (64 MB)  | `open-sse/handlers/chatCore/nonStreamingResponseBody.ts` | Hard cap for a non-streaming upstream response buffered fully into memory. Past this the upstream reader is cancelled and the request fails fast instead of growing an unbounded string until the heap is exhausted.                                                                                                                                                 |
 | `OMNIROUTE_FORWARDING_HEADER_BUDGET_BYTES` | `768`                   | `open-sse/handlers/chatCore/responseHeaders.ts` | Max wire bytes forwarded from upstream response headers. When the budget is exceeded, lower-priority headers (e.g., custom `x-codex-*`, `x-oai-request-id`) are dropped to stay within common reverse-proxy header limits. Set higher to forward more upstream metadata at the cost of larger response header size.                                             |
 | `CORS_ORIGIN`                           | _(unset)_               | `src/server/cors/origins.ts`                    | Legacy single-origin CORS allowlist. Prefer `CORS_ALLOWED_ORIGINS` for new deployments. CORS is only for cross-origin browser API clients; authenticated dashboard writes use same-origin requests plus session-bound CSRF protection instead.                                                                                                                     |
@@ -853,7 +853,6 @@ Reverse-engineered session bridge for hyperagent.com (`src/shared/constants/prov
 
 | Variable                            | Default       | Source File                        | Description                                                                                                                                                                                                                                                                                            |
 | ----------------------------------- | ------------- | ---------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
-| `MODELS_DEV_SYNC_ENABLED`           | `false`       | `src/lib/modelsDevSync.ts`         | Opt-in switch for the models.dev capability sync. Set to anything non-empty it wins over the `modelsDevSyncEnabled` setting (Dashboard > Settings > AI) in either direction, so a deployment can pin the sync on or off without depending on database state surviving a rebuild; unset, it defers to that setting. On for `1`, `true`, `yes` or `on` in any casing; any other value is off.          |
 | `MODELS_DEV_SYNC_INTERVAL`          | `86400` (24h) | `src/lib/modelsDevSync.ts`         | Development-time model catalog sync interval in seconds.                                                                                                                                                                                                                                              |
 | `CONTEXT_WINDOW_RECONCILE_INTERVAL` | `86400` (24h) | `src/lib/contextWindowResolver.ts` | Interval (seconds) for the self-correcting context-window reconciler (5004): pins provider-declared windows from `/models` discovery as `auto:discovery` overrides when they diverge from the catalog. Set to `0` to disable. Reuses already-synced data (no new fetch); never overwrites `manual` overrides. |
 
@@ -1207,17 +1206,6 @@ Provider quota endpoints, network tunnels (Tailscale, Ngrok, MITM debug proxy),
 | `OMNIROUTE_ROTATE_400_THRESHOLD`            | `1`                                                                         | `open-sse/services/rotationConfig.ts`                                     | Number of `400` errors within `OMNIROUTE_ROTATE_400_WINDOW_SECONDS` required before the account is rotated (only consulted when `OMNIROUTE_ROTATE_ON_400=true`).                                                                                                                                                                                                     |
 | `OMNIROUTE_ROTATE_400_WINDOW_SECONDS`       | `120`                                                                       | `open-sse/services/rotationConfig.ts`                                     | Sliding window (seconds) over which `400` errors are counted toward `OMNIROUTE_ROTATE_400_THRESHOLD`.                                                                                                                                                                                                                                                                 |
 
-### Claude Warmup Scheduler
-
-Cron-driven warmup for opted-in Anthropic OAuth connections, so the 5-hour rate-limit window is opened by a trivial scheduled request instead of by the first real one (#8848). The scheduler is off unless `OMNIROUTE_WARMUP_ENABLED` is truthy **and** the connection is flagged in `settings.claudeWarmup.connections`; an empty connection list means nothing is warmed even with the env var on.
-
-| Variable                      | Default                                | Source File                  | Description                                                                                                                                                            |
-| ----------------------------- | -------------------------------------- | ---------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
-| `OMNIROUTE_WARMUP_ENABLED`    | _(unset → off)_                        | `src/lib/warmupScheduler.ts` | Master switch for the warmup scheduler. Accepts `1`/`true`/`yes`/`on` (case-insensitive, trimmed). Any other value, or unset, leaves the scheduler off.                 |
-| `OMNIROUTE_WARMUP_CRON`       | `0 7 * * *`                            | `src/lib/warmupScheduler.ts` | Five-field cron expression for the warmup tick, evaluated in `America/Los_Angeles` (Anthropic's reset timezone) regardless of the host clock.                           |
-| `OMNIROUTE_WARMUP_CONCURRENCY` | `3`                                    | `src/lib/warmupScheduler.ts` | How many connections are warmed in parallel per tick. Clamped to `1`-`10`; a non-numeric value falls back to `3`.                                                       |
-| `OMNIROUTE_WARMUP_MODEL`      | `claude-3-5-haiku-20241022`            | `src/lib/warmupScheduler.ts` | Model used for the warmup request. Override only if the default is unavailable on your plan; pick the cheapest model that still opens the window.                       |
-
 ### Browser-Login VNC Sessions & Data-Dir Alias
 
 Containerized Chromium+VNC used for interactive browser-login credential capture (`/api/vnc-session`), plus a legacy `DATA_DIR` alias. All optional — the VNC defaults target the bundled `omniroute-vnc-chromium:local` image and are only overridden for a custom container image, ports, or lifecycle tuning.
@@ -1287,17 +1275,14 @@ that should be able to run the docs translator.
 Optional add-on gated by the RADAR_ENABLED feature flag (default off — a feature
 flag toggled via Settings/DB, not an env var; see
 [docs/frameworks/RADAR.md](../frameworks/RADAR.md#flag-radar_enabled-default-off)).
-The four variables below are optional overrides used only to point the client at a
-self-hosted or forked feed / supporter-key flow instead of the default OmniRoute
-Radar service. See [docs/frameworks/RADAR.md](../frameworks/RADAR.md) for the full
-module doc.
+Both variables below are optional overrides used only to point the client at a
+self-hosted or forked feed instead of the default OmniRoute Radar feed. See
+[docs/frameworks/RADAR.md](../frameworks/RADAR.md) for the full module doc.
 
-| Variable                       | Default                                          | Source File                   | Description                                                                                     |
-| -------------------------------- | --------------------------------------------------- | ------------------------------ | ------------------------------------------------------------------------------------------------ |
-| `RADAR_FEED_URL`               | `https://radar.omniroute.online`                 | `src/lib/radar/sync.ts`       | Base URL of the Radar feed service. Override to point at a self-hosted or forked feed.          |
-| `RADAR_FEED_PUBKEY`            | _(pinned default key)_                           | `src/lib/radar/pinnedKeys.ts` | Ed25519 public key (base64-DER SPKI or PEM) used to verify feed signatures from a custom feed.   |
-| `RADAR_CONTRIBUTOR_CLAIM_URL`  | `https://radar.omniroute.online/auth/github`     | `src/lib/radar/links.ts`      | URL the "I'm a contributor" dashboard button opens (GitHub OAuth supporter-key claim flow).      |
-| `RADAR_SUPPORTER_PLANS_URL`    | `https://radar.omniroute.online/planos`          | `src/lib/radar/links.ts`      | URL the "Support the project" dashboard button opens (payment/plans page).                       |
+| Variable            | Default                        | Source File                   | Description                                                                                     |
+| -------------------- | ------------------------------- | ------------------------------ | ------------------------------------------------------------------------------------------------ |
+| `RADAR_FEED_URL`    | `https://radar.omniroute.dev`  | `src/lib/radar/sync.ts`       | Base URL of the Radar feed service. Override to point at a self-hosted or forked feed.          |
+| `RADAR_FEED_PUBKEY` | _(pinned default key)_          | `src/lib/radar/pinnedKeys.ts` | Ed25519 public key (base64-DER SPKI or PEM) used to verify feed signatures from a custom feed.   |
 
 ---
 

electron/package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "omniroute-desktop",
-  "version": "3.8.50",
+  "version": "3.8.49",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "omniroute-desktop",
-      "version": "3.8.50",
+      "version": "3.8.49",
       "license": "MIT",
       "dependencies": {
         "electron-updater": "^6.8.9"
@@ -297,6 +297,45 @@
         "url": "https://github.com/sponsors/isaacs"
       }
     },
+    "node_modules/@electron/windows-sign": {
+      "version": "1.2.2",
+      "resolved": "https://registry.npmjs.org/@electron/windows-sign/-/windows-sign-1.2.2.tgz",
+      "integrity": "sha512-dfZeox66AvdPtb2lD8OsIIQh12Tp0GNCRUDfBHIKGpbmopZto2/A8nSpYYLoedPIHpqkeblZ/k8OV0Gy7PYuyQ==",
+      "dev": true,
+      "license": "BSD-2-Clause",
+      "optional": true,
+      "peer": true,
+      "dependencies": {
+        "cross-dirname": "^0.1.0",
+        "debug": "^4.3.4",
+        "fs-extra": "^11.1.1",
+        "minimist": "^1.2.8",
+        "postject": "^1.0.0-alpha.6"
+      },
+      "bin": {
+        "electron-windows-sign": "bin/electron-windows-sign.js"
+      },
+      "engines": {
+        "node": ">=14.14"
+      }
+    },
+    "node_modules/@electron/windows-sign/node_modules/fs-extra": {
+      "version": "11.3.6",
+      "resolved": "https://registry.npmjs.org/fs-extra/-/fs-extra-11.3.6.tgz",
+      "integrity": "sha512-w8ZNZr2mKIc7qeNaQ9AVPT1+iFaI+Avd4xudVOvdDJ8VytREi1Ft5Ih7hd9jjehod8vAM5GMsfQ/TpPf4EyoEA==",
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "peer": true,
+      "dependencies": {
+        "graceful-fs": "^4.2.0",
+        "jsonfile": "^6.0.1",
+        "universalify": "^2.0.0"
+      },
+      "engines": {
+        "node": ">=14.14"
+      }
+    },
     "node_modules/@isaacs/fs-minipass": {
       "version": "4.0.1",
       "resolved": "https://registry.npmjs.org/@isaacs/fs-minipass/-/fs-minipass-4.0.1.tgz",
@@ -1091,6 +1130,15 @@
       "dev": true,
       "license": "MIT"
     },
+    "node_modules/cross-dirname": {
+      "version": "0.1.0",
+      "resolved": "https://registry.npmjs.org/cross-dirname/-/cross-dirname-0.1.0.tgz",
+      "integrity": "sha512-+R08/oI0nl3vfPcqftZRpytksBXDzOUveBq/NBVx0sUp1axwzPQrKinNx5yd5sxPu8j1wIy8AfnVQ+5eFdha6Q==",
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "peer": true
+    },
     "node_modules/cross-spawn": {
       "version": "7.0.6",
       "resolved": "https://registry.npmjs.org/cross-spawn/-/cross-spawn-7.0.6.tgz",
@@ -1411,6 +1459,19 @@
         "node": ">=14.0.0"
       }
     },
+    "node_modules/electron-builder-squirrel-windows": {
+      "version": "26.15.3",
+      "resolved": "https://registry.npmjs.org/electron-builder-squirrel-windows/-/electron-builder-squirrel-windows-26.15.3.tgz",
+      "integrity": "sha512-Jc19XPV9y9+2bAdZPkXuVNGNIEFBq9poHC61l8Kv6FdK7DRG3+Ic0rerC0DXOaeHNz8yW0fg/JnF8GQROOF5MA==",
+      "dev": true,
+      "license": "MIT",
+      "peer": true,
+      "dependencies": {
+        "app-builder-lib": "26.15.3",
+        "builder-util": "26.15.3",
+        "electron-winstaller": "5.4.0"
+      }
+    },
     "node_modules/electron-publish": {
       "version": "26.15.3",
       "resolved": "https://registry.npmjs.org/electron-publish/-/electron-publish-26.15.3.tgz",
@@ -1445,6 +1506,66 @@
         "tiny-typed-emitter": "^2.1.0"
       }
     },
+    "node_modules/electron-winstaller": {
+      "version": "5.4.0",
+      "resolved": "https://registry.npmjs.org/electron-winstaller/-/electron-winstaller-5.4.0.tgz",
+      "integrity": "sha512-bO3y10YikuUwUuDUQRM4KfwNkKhnpVO7IPdbsrejwN9/AABJzzTQ4GeHwyzNSrVO+tEH3/Np255a3sVZpZDjvg==",
+      "dev": true,
+      "hasInstallScript": true,
+      "license": "MIT",
+      "peer": true,
+      "dependencies": {
+        "@electron/asar": "^3.2.1",
+        "debug": "^4.1.1",
+        "fs-extra": "^7.0.1",
+        "lodash": "^4.17.21",
+        "temp": "^0.9.0"
+      },
+      "engines": {
+        "node": ">=8.0.0"
+      },
+      "optionalDependencies": {
+        "@electron/windows-sign": "^1.1.2"
+      }
+    },
+    "node_modules/electron-winstaller/node_modules/fs-extra": {
+      "version": "7.0.1",
+      "resolved": "https://registry.npmjs.org/fs-extra/-/fs-extra-7.0.1.tgz",
+      "integrity": "sha512-YJDaCJZEnBmcbw13fvdAM9AwNOJwOzrE4pqMqBq5nFiEqXUqHwlK4B+3pUw6JNvfSPtX05xFHtYy/1ni01eGCw==",
+      "dev": true,
+      "license": "MIT",
+      "peer": true,
+      "dependencies": {
+        "graceful-fs": "^4.1.2",
+        "jsonfile": "^4.0.0",
+        "universalify": "^0.1.0"
+      },
+      "engines": {
+        "node": ">=6 <7 || >=8"
+      }
+    },
+    "node_modules/electron-winstaller/node_modules/jsonfile": {
+      "version": "4.0.0",
+      "resolved": "https://registry.npmjs.org/jsonfile/-/jsonfile-4.0.0.tgz",
+      "integrity": "sha512-m6F1R3z8jjlf2imQHS2Qez5sjKWQzbuuhuJ/FKYFRZvPE3PuHcSMVZzfsLhGVOkfd20obL5SWEBew5ShlquNxg==",
+      "dev": true,
+      "license": "MIT",
+      "peer": true,
+      "optionalDependencies": {
+        "graceful-fs": "^4.1.6"
+      }
+    },
+    "node_modules/electron-winstaller/node_modules/universalify": {
+      "version": "0.1.2",
+      "resolved": "https://registry.npmjs.org/universalify/-/universalify-0.1.2.tgz",
+      "integrity": "sha512-rBJeI5CXAlmy1pV+617WB9J63U6XcazHHF2f2dbJix4XzpUF0RS3Zbj0FGIOCAva5P/d/GBOYaACQ1w+0azUkg==",
+      "dev": true,
+      "license": "MIT",
+      "peer": true,
+      "engines": {
+        "node": ">= 4.0.0"
+      }
+    },
     "node_modules/emoji-regex": {
       "version": "8.0.0",
       "resolved": "https://registry.npmjs.org/emoji-regex/-/emoji-regex-8.0.0.tgz",
@@ -1575,9 +1696,9 @@
       "license": "MIT"
     },
     "node_modules/fast-uri": {
-      "version": "3.1.5",
-      "resolved": "https://registry.npmjs.org/fast-uri/-/fast-uri-3.1.5.tgz",
-      "integrity": "sha512-gHwA1O9LDIcKunMKhObS/HimwtehO1nPUECKAu5TpKgaO19fcWEl4bliWe1jWxVFvIXztJjjQ4L8XQ1EU9f7Jw==",
+      "version": "3.1.4",
+      "resolved": "https://registry.npmjs.org/fast-uri/-/fast-uri-3.1.4.tgz",
+      "integrity": "sha512-8JnbkQ4juDyvYs4mgFGQqg4yCYtFDtUtmp2QIQq11ZZe5CFQ5wcqm1rqDgAh/QdMySuBnPzMUiJUNZG5N/AiQw==",
       "dev": true,
       "funding": [
         {
@@ -2359,6 +2480,20 @@
         "node": ">= 18"
       }
     },
+    "node_modules/mkdirp": {
+      "version": "0.5.6",
+      "resolved": "https://registry.npmjs.org/mkdirp/-/mkdirp-0.5.6.tgz",
+      "integrity": "sha512-FP+p8RB8OWpF3YZBCrP5gtADmtXApB5AMLn+vdyA+PyxCjrCs00mjyUozssO33cwDeT3wNGdLxJ5M//YqtHAJw==",
+      "dev": true,
+      "license": "MIT",
+      "peer": true,
+      "dependencies": {
+        "minimist": "^1.2.6"
+      },
+      "bin": {
+        "mkdirp": "bin/cmd.js"
+      }
+    },
     "node_modules/ms": {
       "version": "2.1.3",
       "resolved": "https://registry.npmjs.org/ms/-/ms-2.1.3.tgz",
@@ -2622,6 +2757,36 @@
         "node": ">=18"
       }
     },
+    "node_modules/postject": {
+      "version": "1.0.0-alpha.6",
+      "resolved": "https://registry.npmjs.org/postject/-/postject-1.0.0-alpha.6.tgz",
+      "integrity": "sha512-b9Eb8h2eVqNE8edvKdwqkrY6O7kAwmI8kcnBv1NScolYJbo59XUF0noFq+lxbC1yN20bmC0WBEbDC5H/7ASb0A==",
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "peer": true,
+      "dependencies": {
+        "commander": "^9.4.0"
+      },
+      "bin": {
+        "postject": "dist/cli.js"
+      },
+      "engines": {
+        "node": ">=14.0.0"
+      }
+    },
+    "node_modules/postject/node_modules/commander": {
+      "version": "9.5.0",
+      "resolved": "https://registry.npmjs.org/commander/-/commander-9.5.0.tgz",
+      "integrity": "sha512-KRs7WVDKg86PWiuAqhDrAQnTXZKraVcCc6vFdL14qrZ/DcWwuRo7VoiYXalXO7S5GKpqYiVEwCbgFDfxNHKJBQ==",
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "peer": true,
+      "engines": {
+        "node": "^12.20.0 || >=14"
+      }
+    },
     "node_modules/proc-log": {
       "version": "6.1.0",
       "resolved": "https://registry.npmjs.org/proc-log/-/proc-log-6.1.0.tgz",
@@ -2816,6 +2981,21 @@
         "node": ">= 4"
       }
     },
+    "node_modules/rimraf": {
+      "version": "2.6.3",
+      "resolved": "https://registry.npmjs.org/rimraf/-/rimraf-2.6.3.tgz",
+      "integrity": "sha512-mwqeW5XsA2qAejG46gYdENaxXjx9onRNCfn7L0duuP4hCuTIi/QO7PDK07KJfp1d+izWPrzEJDcSqBa0OZQriA==",
+      "deprecated": "Rimraf versions prior to v4 are no longer supported",
+      "dev": true,
+      "license": "ISC",
+      "peer": true,
+      "dependencies": {
+        "glob": "^7.1.3"
+      },
+      "bin": {
+        "rimraf": "bin.js"
+      }
+    },
     "node_modules/roarr": {
       "version": "2.15.4",
       "resolved": "https://registry.npmjs.org/roarr/-/roarr-2.15.4.tgz",
@@ -3071,6 +3251,21 @@
         "node": ">=18"
       }
     },
+    "node_modules/temp": {
+      "version": "0.9.4",
+      "resolved": "https://registry.npmjs.org/temp/-/temp-0.9.4.tgz",
+      "integrity": "sha512-yYrrsWnrXMcdsnu/7YMYAofM1ktpL5By7vZhf15CrXijWWrEYZks5AXBudalfSWJLlnen/QUJUB5aoB0kqZUGA==",
+      "dev": true,
+      "license": "MIT",
+      "peer": true,
+      "dependencies": {
+        "mkdirp": "^0.5.1",
+        "rimraf": "~2.6.2"
+      },
+      "engines": {
+        "node": ">=6.0.0"
+      }
+    },
     "node_modules/temp-file": {
       "version": "3.4.0",
       "resolved": "https://registry.npmjs.org/temp-file/-/temp-file-3.4.0.tgz",
@@ -3177,9 +3372,9 @@
       }
     },
     "node_modules/undici": {
-      "version": "7.29.0",
-      "resolved": "https://registry.npmjs.org/undici/-/undici-7.29.0.tgz",
-      "integrity": "sha512-IDxfleLmmbSskfWSUATiN1nfn2rDuvnMOqb5CWR92iIfojA0Ud+ulOAAEQ57LPr9rWmsreUyf5lwyao+7GNNVw==",
+      "version": "7.28.0",
+      "resolved": "https://registry.npmjs.org/undici/-/undici-7.28.0.tgz",
+      "integrity": "sha512-cRZYrTDwWznlnRiPjggAGxZXanty6M8RV1ff8Wm4LWXBp7/IG8v5DnOm74DtUBp9OONpK75YlPnIjQqX0dBDtA==",
       "dev": true,
       "license": "MIT",
       "engines": {

electron/package.json

@@ -37,7 +37,7 @@
     "plist": "^4.0.0",
     "form-data": "^4.0.6",
     "js-yaml": "^4.2.0",
-    "undici": "^7.29.0"
+    "undici": "^7.28.0"
   },
   "build": {
     "appId": "online.omniroute.desktop",

open-sse/config/audioRegistry.ts

@@ -145,19 +145,6 @@ export const AUDIO_TRANSCRIPTION_PROVIDERS: Record<string, AudioProvider> = {
     ],
   },
 
-  soniox: {
-    id: "soniox",
-    baseUrl: "https://api.soniox.com/v1/transcriptions",
-    authType: "apikey",
-    authHeader: "bearer",
-    async: true,
-    format: "soniox",
-    models: [
-      { id: "stt-async-v5", name: "Soniox STT Async v5" },
-      { id: "stt-async-v4", name: "Soniox STT Async v4" },
-    ],
-  },
-
   nvidia: {
     id: "nvidia",
     baseUrl: "https://integrate.api.nvidia.com/v1/audio/transcriptions",
@@ -333,15 +320,6 @@ export const AUDIO_SPEECH_PROVIDERS: Record<string, AudioProvider> = {
     ],
   },
 
-  soniox: {
-    id: "soniox",
-    baseUrl: "https://tts-rt.soniox.com/tts",
-    authType: "apikey",
-    authHeader: "bearer",
-    format: "soniox-tts",
-    models: [{ id: "tts-rt-v1", name: "Soniox TTS RT v1" }],
-  },
-
   elevenlabs: {
     id: "elevenlabs",
     baseUrl: "https://api.elevenlabs.io/v1/text-to-speech",

open-sse/config/providers/registry/opencode/zen/index.ts

@@ -85,13 +85,14 @@ export const opencode_zenProvider: RegistryEntry = {
     { id: "qwen3.6-plus", name: "Qwen3.6 Plus", targetFormat: "claude", supportsVision: false },
 
     // ── Free Tier ──────────────────────────────────────────────
-    // #6998 (2026-07-14): upstream free tier rotated — minimax-m2.5-free,
-    // nemotron-3-super-free and qwen3.6-plus-free were delisted (401). Replaced
-    // by the 4 entries below with upstream-verified limits.
     { id: "deepseek-v4-flash-free", name: "DeepSeek V4 Flash Free", supportsReasoning: true },
-    { id: "mimo-v2.5-free", name: "MiMo V2.5 Free", contextLength: 200000 },
-    { id: "hy3-free", name: "HY3 Free", contextLength: 200000 },
-    { id: "nemotron-3-ultra-free", name: "Nemotron 3 Ultra Free", contextLength: 1000000 },
-    { id: "north-mini-code-free", name: "North Mini Code Free", contextLength: 200000 },
+    { id: "minimax-m2.5-free", name: "MiniMax M2.5 Free", contextLength: 204800 },
+    { id: "nemotron-3-super-free", name: "Nemotron 3 Super Free", contextLength: 1000000 },
+    {
+      id: "qwen3.6-plus-free",
+      name: "Qwen3.6 Plus Free",
+      targetFormat: "claude",
+      contextLength: 200000,
+    },
   ],
 };

open-sse/config/providers/registry/openrouter/index.ts

@@ -13,12 +13,5 @@ export const openrouterProvider: RegistryEntry = {
     "HTTP-Referer": "https://endpoint-proxy.local",
     "X-Title": "Endpoint Proxy",
   },
-  // OpenRouter multiplexes hundreds of independent upstream models behind one
-  // connection/API key — without this flag, hasPerModelQuota() (accountFallback.ts)
-  // falls through to connection-wide cooldown on any model-specific failure (e.g. a
-  // 404 "No endpoints found" for one dead/renamed model), poisoning every OTHER
-  // OpenRouter model on the same connection for the cooldown window and surfacing
-  // that first model's stale error message on their unrelated requests.
-  passthroughModels: true,
   models: [{ id: "auto", name: "Auto (Best Available)" }],
 };

open-sse/config/providers/shared.ts

@@ -271,14 +271,16 @@ export const GPT_5_6_API_CAPABILITIES = {
   maxOutputTokens: 128000,
 } as const;
 
+// Codex's live catalog reports a 272K input context window for GPT-5.6.
+// Keep the input and output limits explicit for catalog consumers that expose them separately.
 export const GPT_5_6_CODEX_CAPABILITIES = {
   targetFormat: "openai-responses",
   toolCalling: true,
   supportsReasoning: true,
   supportsVision: true,
   supportsXHighEffort: true,
-  contextLength: 1050000,
-  maxInputTokens: 922000,
+  contextLength: 272000,
+  maxInputTokens: 272000,
   maxOutputTokens: 128000,
 } as const;
 

open-sse/executors/antigravity.ts

@@ -23,11 +23,6 @@ import {
 import { persistCreditBalance, getAllPersistedCreditBalances } from "@/lib/db/creditBalance";
 import { setConnectionRateLimitUntil } from "@/lib/db/providers";
 import { getMitmAlias } from "@/lib/db/models";
-import {
-  MAX_ANTIGRAVITY_OUTPUT_TOKENS,
-  resolveAntigravityOutputCap,
-} from "./antigravityOutputCap.ts";
-export { MAX_ANTIGRAVITY_OUTPUT_TOKENS } from "./antigravityOutputCap.ts";
 import { ensureAntigravityProjectAssigned } from "../services/antigravityProjectBootstrap.ts";
 import { persistDiscoveredAntigravityProjectId } from "../services/antigravityProjectPersist.ts";
 import {
@@ -284,10 +279,18 @@ async function cleanModelName(model: string, modelIdOverride?: string): Promise<
   return clean;
 }
 
-function applyAntigravityGenerationDefaults(
-  request: Record<string, unknown>,
-  modelId?: string | null
-): void {
+/**
+ * Hard ceiling on `generationConfig.maxOutputTokens` for Antigravity Cloud Code.
+ *
+ * Ports decolua/9router#779 (lukmanfauzie): VS Code GitHub Copilot Chat in
+ * Agent mode regularly requests 32K–65K output tokens, which the Antigravity
+ * backend rejects with HTTP 400 "Invalid Argument". 16384 matches the
+ * upstream-accepted ceiling confirmed via successful 200 OK runs with
+ * claude-sonnet-4-6 and gemini-pro-agent across both Ask and Agent modes.
+ */
+export const MAX_ANTIGRAVITY_OUTPUT_TOKENS = 16384;
+
+function applyAntigravityGenerationDefaults(request: Record<string, unknown>): void {
   const generationConfig =
     request.generationConfig && typeof request.generationConfig === "object"
       ? (request.generationConfig as Record<string, unknown>)
@@ -319,10 +322,9 @@ function applyAntigravityGenerationDefaults(
   // (32K–65K) that trigger upstream 400 "Invalid Argument". Clamp silently
   // — the cap is provider-driven, not client-driven, and only matters when
   // the request would otherwise be rejected outright.
-  const cap = resolveAntigravityOutputCap(modelId);
   const finalMax = Number(generationConfig.maxOutputTokens);
-  if (Number.isFinite(finalMax) && finalMax > cap) {
-    generationConfig.maxOutputTokens = cap;
+  if (Number.isFinite(finalMax) && finalMax > MAX_ANTIGRAVITY_OUTPUT_TOKENS) {
+    generationConfig.maxOutputTokens = MAX_ANTIGRAVITY_OUTPUT_TOKENS;
   }
 
   request.generationConfig = generationConfig;
@@ -664,7 +666,7 @@ export class AntigravityExecutor extends BaseExecutor {
         )
       : rawTransformedRequest;
 
-    applyAntigravityGenerationDefaults(transformedRequest, upstreamModel);
+    applyAntigravityGenerationDefaults(transformedRequest);
 
     const {
       project: _project,

open-sse/executors/antigravityOutputCap.ts

@@ -1,52 +0,0 @@
-import { getExplicitModelOutputCap } from "@/lib/modelCapabilities";
-
-/**
- * Fallback ceiling on `generationConfig.maxOutputTokens` for Antigravity
- * Cloud Code, used when the model is unknown to the catalogue.
- *
- * Ports decolua/9router#779 (lukmanfauzie): VS Code GitHub Copilot Chat in
- * Agent mode regularly requests 32K–65K output tokens, which the Antigravity
- * backend rejects with HTTP 400 "Invalid Argument". 16384 was the ceiling
- * confirmed safe at the time, via successful 200 OK runs with
- * claude-sonnet-4-6 and gemini-pro-agent across both Ask and Agent modes.
- *
- * Both of those models are catalogue-known today, so neither one reaches this
- * constant anymore: they get their own declared limit via
- * `resolveAntigravityOutputCap` (65536 and 65535, respectively). The higher
- * limit holds against the live upstream. A gemini-3.6-flash-high request came
- * back with completion_tokens 16754 and finish_reason "stop", which exceeds
- * 16384 on its own and so cannot be an artifact of thinking-token accounting.
- *
- * Note also that #779 was reported against Copilot Chat in Agent mode, a path
- * that does not reach this executor, so 16384 arrived with that port rather
- * than from a limit measured here. Beware of re-deriving it from a running
- * instance: the clamp below rewrites maxOutputTokens before the request
- * leaves, so a build still carrying a low constant measures its own clamp and
- * reports it as an upstream ceiling.
- */
-export const MAX_ANTIGRAVITY_OUTPUT_TOKENS = 16384;
-
-/**
- * The output ceiling this specific model accepts, or the conservative
- * fallback above when the id is not in the catalogue.
- *
- * The declared limits are not uniform: most Antigravity models publish
- * 65535 or 65536, but gpt-oss-120b-medium publishes 32768. A single global
- * ceiling either starves the first group or lets an oversized request
- * through to the second, so the number has to come from the model.
- */
-export function resolveAntigravityOutputCap(modelId: string | null | undefined): number {
-  const id = typeof modelId === "string" ? modelId.trim() : "";
-  if (!id) return MAX_ANTIGRAVITY_OUTPUT_TOKENS;
-  try {
-    const declared = getExplicitModelOutputCap({ provider: "antigravity", model: id });
-    return typeof declared === "number" && Number.isFinite(declared) && declared > 0
-      ? declared
-      : MAX_ANTIGRAVITY_OUTPUT_TOKENS;
-  } catch {
-    // DB not available (build phase, transient error) -- fall through to the
-    // conservative fallback, the same guard cleanModelName uses above for
-    // its own MITM alias lookup.
-    return MAX_ANTIGRAVITY_OUTPUT_TOKENS;
-  }
-}

open-sse/executors/azure-openai.ts

@@ -28,11 +28,7 @@ export class AzureOpenAIExecutor extends DefaultExecutor {
     void urlIndex;
 
     const providerSpecificData = credentials?.providerSpecificData || {};
-    const baseUrl = normalizeAzureBaseUrl(
-      typeof providerSpecificData.baseUrl === "string"
-        ? providerSpecificData.baseUrl
-        : this.config.baseUrl
-    );
+    const baseUrl = normalizeAzureBaseUrl(providerSpecificData.baseUrl || this.config.baseUrl);
     const apiVersion =
       typeof providerSpecificData.apiVersion === "string" && providerSpecificData.apiVersion.trim()
         ? providerSpecificData.apiVersion.trim()

open-sse/executors/base.ts

@@ -234,11 +234,71 @@ export function mergeAbortSignals(primary: AbortSignal, secondary: AbortSignal):
   return controller.signal;
 }
 
-import {
-  hasActiveClaudeThinking,
-  readNestedThinkingBudget,
-  clampNestedThinkingBudget,
-} from "../utils/thinkingBudget.ts";
+function hasActiveClaudeThinking(body: Record<string, unknown>): boolean {
+  const thinking = body.thinking as Record<string, unknown> | undefined;
+  return thinking?.type === "enabled" || thinking?.type === "adaptive";
+}
+
+/**
+ * Collect every `thinkingConfig` object in a transformed request body that holds
+ * a thinking budget, wherever the provider's envelope nests it:
+ *   - body.generationConfig.thinkingConfig            (native Gemini / openai→gemini)
+ *   - body.request.generationConfig.thinkingConfig    (Antigravity Cloud Code envelope)
+ * Returns only objects that actually carry a `thinkingBudget`/`thinking_budget`
+ * field — a request without thinking config is never mutated.
+ */
+function collectThinkingConfigs(body: unknown): Array<Record<string, unknown>> {
+  if (!body || typeof body !== "object") return [];
+  const root = body as Record<string, unknown>;
+  const configs: Array<Record<string, unknown>> = [];
+  const envelopes: unknown[] = [
+    root.generationConfig,
+    (root.request as Record<string, unknown> | undefined)?.generationConfig,
+  ];
+  for (const env of envelopes) {
+    if (!env || typeof env !== "object") continue;
+    const tc = (env as Record<string, unknown>).thinkingConfig;
+    if (tc && typeof tc === "object") {
+      const tcr = tc as Record<string, unknown>;
+      if ("thinkingBudget" in tcr || "thinking_budget" in tcr) configs.push(tcr);
+    }
+  }
+  return configs;
+}
+
+/**
+ * Read the first thinking budget found in the body (any supported nest / naming).
+ * Returns null when the body carries no readable numeric budget.
+ */
+function readNestedThinkingBudget(body: unknown): number | null {
+  for (const tc of collectThinkingConfigs(body)) {
+    const raw = tc.thinkingBudget ?? tc.thinking_budget;
+    const n = Number(raw);
+    if (Number.isFinite(n)) return n;
+  }
+  return null;
+}
+
+/**
+ * Clamp every thinking budget in the body down to `max` (only lowers; never
+ * raises a budget already below max). Mutates in place. Returns true when at
+ * least one budget was actually lowered (i.e. a retry would send a different
+ * body) — false means the 400 was not caused by an over-max budget we hold, so
+ * retrying would resend an identical body and loop.
+ */
+function clampNestedThinkingBudget(body: unknown, max: number): boolean {
+  let changed = false;
+  for (const tc of collectThinkingConfigs(body)) {
+    for (const key of ["thinkingBudget", "thinking_budget"] as const) {
+      const n = Number(tc[key]);
+      if (Number.isFinite(n) && n > max) {
+        tc[key] = max;
+        changed = true;
+      }
+    }
+  }
+  return changed;
+}
 
 /**
  * Strip the OmniRoute provider prefix from tool model fields (e.g.

open-sse/executors/base/reasoningEffort.ts

@@ -151,8 +151,7 @@ export function supportsMaxEffortForProvider(provider: string, model: string): b
   // Ollama Cloud also accepts literal max (for example GLM 5.2 supports
   // low|medium|high|max|none) and rejects xhigh.
   const isOpencodeGoDeepSeek =
-    (provider === "opencode-go" || provider === "opencode-zen") &&
-    model.toLowerCase().includes("deepseek");
+    provider === "opencode-go" && model.toLowerCase().includes("deepseek");
   const isOllamaCloud = provider === "ollama-cloud";
   const isMoonshotK3 =
     (provider === "moonshot" || provider === "kimi") && /^kimi-k3(?:$|-)/i.test(model);

open-sse/executors/chatgpt-web.ts

@@ -2816,14 +2816,11 @@ export class ChatGptWebExecutor extends BaseExecutor {
       };
     }
 
-    // Tool-call emulation (#5240, #7679): inject a `<tool>` contract when tools
-    // are present; parsed back on the response side. Hardened for thinking models.
-    const resolvedModel = resolveChatGptModel(model, body, credentials.providerSpecificData);
-    const modelSlug = resolvedModel.slug;
+    // Tool-call emulation (#5240): inject a `<tool>` contract when `tools` are
+    // present; parsed back on the response side. Mirrors qwen-web/perplexity-web.
     const { hasTools, requestedTools, effectiveMessages } = prepareToolMessages(
       (body || {}) as Record<string, unknown>,
-      messages as Array<{ role: string; content: unknown }>,
-      { hardened: isThinkingCapableModel(model, modelSlug) }
+      messages as Array<{ role: string; content: unknown }>
     );
 
     if (!credentials.apiKey) {
@@ -2921,9 +2918,12 @@ export class ChatGptWebExecutor extends BaseExecutor {
       log
     );
 
-    // 2a''. Apply thinking-effort preference for thinking models.
-    // Dedicated thinking models mirror the browser's user-config PATCH;
-    // GPT-5.5 Pro effort is sent with the conversation body.
+    // 2a''. Resolve model + effort and apply thinking-effort preference for
+    // thinking-capable models. Dedicated thinking models mirror the browser's
+    // user-config PATCH; GPT-5.5 Pro sends the effort with the conversation
+    // body because the Pro standard/extended budget is part of that turn.
+    const resolvedModel = resolveChatGptModel(model, body, credentials.providerSpecificData);
+    const modelSlug = resolvedModel.slug;
     const requestedEffort = resolvedModel.effort;
     if (requestedEffort && isThinkingCapableModel(model, modelSlug)) {
       await setUserThinkingEffort(

open-sse/executors/commandCode.ts

@@ -461,31 +461,14 @@ function applyEventToAggregateOrThrow(event: JsonRecord, state: AggregateState):
 function usageFromCommandCode(usage: JsonRecord | null) {
   if (!usage) return undefined;
   const details = isRecord(usage.inputTokenDetails) ? usage.inputTokenDetails : {};
-  const cacheRead = numberValue(details.cacheReadTokens) || 0;
-  const noCache = numberValue(details.noCacheTokens) || 0;
-  // Command Code's totalUsage.inputTokens is the FULL prompt total and already
-  // includes the cached portion (noCacheTokens + cacheReadTokens = inputTokens),
-  // so we must NOT add cacheRead back — that would double-count. There is no
-  // cache-write field in the upstream payload, so cache creation stays unset.
-  const inputTokens = numberValue(usage.inputTokens) || 0;
-  const prompt = inputTokens;
+  const prompt =
+    (numberValue(usage.inputTokens) || 0) + (numberValue(details.cacheReadTokens) || 0);
   const completion = numberValue(usage.outputTokens) || 0;
-  const result: JsonRecord = {
+  return {
     prompt_tokens: prompt,
     completion_tokens: completion,
     total_tokens: prompt + completion,
   };
-  // Surface the cache breakdown as informational fields so logUsage prints
-  // `| cache_read=X | no_cache=Y` and appendRequestLog persists them. These are
-  // NOT added to prompt_tokens (already included) — metering stays accurate.
-  if (cacheRead > 0) result.cache_read_input_tokens = cacheRead;
-  if (noCache > 0) result.no_cache_tokens = noCache;
-  // Keep reasoning_token_details (reasoningTokens) when present so stream.ts's
-  // extractUsage can surface it as reasoning_tokens.
-  const reasoningDetails = isRecord(usage.reasoningTokenDetails) ? usage.reasoningTokenDetails : {};
-  const reasoning = numberValue(reasoningDetails.reasoningTokens);
-  if (reasoning !== undefined && reasoning > 0) result.reasoning_tokens = reasoning;
-  return result;
 }
 
 function createStreamResponse(
@@ -566,22 +549,6 @@ function createStreamResponse(
             state.finishReason = mapFinishReason(event.finishReason);
             state.usage = isRecord(event.totalUsage) ? event.totalUsage : null;
             controller.enqueue(sse(chatCompletionChunk(id, model, {}, state.finishReason)));
-            // Emit a standards-compliant usage-only chunk (choices: []) before
-            // [DONE] when upstream reported usage. stream.ts's extractUsage
-            // recognizes this shape (see stream.ts:1661) and logs the ACTUAL
-            // token counts (in/out/cache_read/no_cache) instead of estimates.
-            const usagePayload = usageFromCommandCode(state.usage);
-            if (usagePayload) {
-              controller.enqueue(
-                sse({
-                  id,
-                  object: "chat.completion.chunk",
-                  model,
-                  usage: usagePayload,
-                  choices: [],
-                })
-              );
-            }
             controller.enqueue(encoder.encode("data: [DONE]\n\n"));
             closed = true;
             controller.close();

open-sse/executors/cursor.ts

@@ -1219,9 +1219,6 @@ export class CursorExecutor extends BaseExecutor {
 
     if (isToolFollowUp) {
       session = cursorSessionManager.acquire(conversationId);
-      // #9029: content-based session match when client lacks conversation_id.
-      if (!session && !body.conversation_id) session = cursorSessionManager.findByToolCallIds(
-        messages.filter(m => m.role === "tool" && m.tool_call_id).map(m => m.tool_call_id!));
     }
 
     if (session) {

open-sse/executors/lmarena/response.ts

@@ -165,7 +165,7 @@ function baseChunk(model: string) {
 }
 
 function enqueueSse(controller: ReadableStreamDefaultController, chunk: Record<string, unknown>) {
-  controller.enqueue(new TextEncoder().encode(`data: ${JSON.stringify(chunk)}\n\n`));
+  controller.enqueue(`data: ${JSON.stringify(chunk)}\n\n`);
 }
 
 function emitStopAndDone(controller: ReadableStreamDefaultController, model: string) {
@@ -173,7 +173,7 @@ function emitStopAndDone(controller: ReadableStreamDefaultController, model: str
     ...baseChunk(model),
     choices: [{ index: 0, delta: {}, finish_reason: "stop" }],
   });
-  controller.enqueue(new TextEncoder().encode("data: [DONE]\n\n"));
+  controller.enqueue("data: [DONE]\n\n");
   controller.close();
 }
 
@@ -213,7 +213,7 @@ export function createOpenAIArenaStream(opts: {
   model: string;
   signal?: AbortSignal;
   log?: { error?: (scope: string, msg: string) => void };
-}): ReadableStream<Uint8Array> {
+}): ReadableStream<Uint8Array | string> {
   const { reader, model, signal, log } = opts;
   const decoder = new TextDecoder();
   let buffer = "";

open-sse/executors/opencode.ts

@@ -40,31 +40,6 @@ const OPENCODE_COOLDOWN_MAX_MS = 60_000;
 
 const EFFORT_LEVELS = ["low", "medium", "high", "max"] as const;
 
-/**
- * Models that work WITHOUT any API key on the free/noauth opencode tier.
- *
- * The upstream free tier rotates frequently — when a `-free` suffix model is
- * delisted upstream, the upstream returns "Model X is not supported" (a separate
- * issue from this gate). The set is defined by two data sources:
- *
- *   1. **Known free models** — models explicitly listed in the noauth
- *      `opencode` provider registry (`open-sse/config/providers/registry/opencode/index.ts`).
- *      These are the canonical free models. `deepseek-v4-flash-free` appears in both
- *      the noauth AND the zen registry (it is free on both tiers).
- *   2. **`-free` suffix** — any model whose id ends in `-free`. This automatically
- *      covers upstream free-tier additions without a code deploy.
- *
- * For `opencode-go`, there is no free tier — ALL models require an API key.
- */
-const OPENCODE_FREE_MODELS = new Set([
-  "big-pickle",
-  "deepseek-v4-flash-free",
-  "mimo-v2.5-free",
-  "hy3-free",
-  "nemotron-3-ultra-free",
-  "north-mini-code-free",
-]);
-
 /**
  * Models on opencode-go that support effort-tier aliases. Each entry maps the
  * canonical base id to the set of effort suffixes the upstream supports.
@@ -111,31 +86,7 @@ export function parseEffortLevel(model: string): { baseModel: string; effort: st
   return null;
 }
 
-/**
- * Determine whether a model requires an API key on the given opencode provider.
- *
- * - `opencode-go`: ALL models require a key (no free tier).
- * - `opencode` / `opencode-zen`: premium = any model NOT in the free set (known
- *   free models OR ending in `-free`).
- * - Unknown models are assumed premium (fail-safe).
- */
-export function isPremiumOpencodeModel(model: string, provider: string): boolean {
-  // opencode-go has no free tier — every model requires a key.
-  if (provider === "opencode-go") return true;
-
-  // Models ending in `-free` are always free on the noauth/zen tier.
-  if (model.endsWith("-free")) return false;
-
-  // Check the known free model catalog.
-  return !OPENCODE_FREE_MODELS.has(model);
-}
-
 export class OpencodeExecutor extends BaseExecutor {
-  /** Delegates to `isPremiumOpencodeModel`. Exported for testability. */
-  static isPremiumModel(model: string, provider: string): boolean {
-    return isPremiumOpencodeModel(model, provider);
-  }
-
   _requestFormat: string | null = null;
 
   /**
@@ -230,34 +181,6 @@ export class OpencodeExecutor extends BaseExecutor {
 
   async execute(input: ExecuteInput) {
     this._requestFormat = getModelTargetFormat(this.provider, input.model) || "openai";
-
-    // #8681: Gate premium opencode models behind a usable API key.
-    // When the connection is keyless (no apiKey, no accessToken) and the model
-    // is a premium model (not on the free tier), return a clear 402 error
-    // instead of proxying the raw upstream 401 "Missing API key" response.
-    const creds = input.credentials;
-    const isKeyless =
-      !creds?.apiKey && !creds?.accessToken && !creds?.providerSpecificData?.extraApiKeys;
-    if (isKeyless && isPremiumOpencodeModel(input.model, this.provider)) {
-      const bodyJson = JSON.stringify({
-        error: {
-          message:
-            "This model requires an opencode API key — add one in Settings → Providers.",
-          type: "invalid_request_error",
-          code: "premium_model_requires_key",
-        },
-      });
-      return {
-        response: new Response(bodyJson, {
-          status: 402,
-          headers: { "Content-Type": "application/json" },
-        }),
-        url: "",
-        headers: {} as Record<string, string>,
-        transformedBody: null,
-      };
-    }
-
     try {
       this.syncAccountsFromCredentials(input.credentials);
 

open-sse/executors/qoder.ts

@@ -372,16 +372,8 @@ export class QoderExecutor extends BaseExecutor {
 
     const { text, isError, errorMessage } = parseQoderCliResult(run.stdout);
     if (isError) {
-      // When qodercli exits 0 but returns is_error=true with an empty result,
-      // the real upstream error is almost always on stderr. Surface it instead
-      // of the generic "qodercli returned an error" fallback (#9319).
-      let effectiveError = errorMessage;
-      if (errorMessage === "qodercli returned an error" && run.stderr.trim()) {
-        const stderrTrimmed = run.stderr.trim().slice(0, 300);
-        effectiveError = `qodercli returned an error: ${stderrTrimmed}`;
-      }
       return {
-        response: createQoderErrorResponse(parseQoderCliFailure(effectiveError)),
+        response: createQoderErrorResponse(parseQoderCliFailure(errorMessage)),
         url,
         headers: {},
         transformedBody: body,

open-sse/executors/qwen-web.ts

@@ -47,8 +47,8 @@ const BX_UMIDTOKEN_FALLBACK = "T2gA0000000000000000000000000000000000000000";
 // header the upstream returns HTTP 200 with `{"success":false,"data":{"code":"Bad_Request"}}`
 // for every completion request, even with a valid session. The version string is
 // the SPA build identifier shipped in the React client's `version` request header.
-// Pinned from a live capture (2026-08); bump if Qwen ships a breaking change.
-const QWEN_SPA_VERSION = "0.2.81";
+// Pinned from a live capture (2026-07); bump if Qwen ships a breaking change.
+const QWEN_SPA_VERSION = "0.2.66";
 
 const MODEL_ALIASES: Record<string, string> = {
   // Legacy OmniRoute ids → current upstream catalog (GET /api/models).

open-sse/executors/vertex.ts

@@ -138,149 +138,13 @@ function isPartnerModel(model: string) {
   return [...PARTNER_MODELS].some((prefix) => normalizedModel.startsWith(prefix));
 }
 
-// Anthropic models need their own branch: they use Vertex's native Anthropic Messages API
-// (publishers/anthropic/.../rawPredict), not the generic OpenAI-compatible partner endpoint the
-// other PARTNER_MODELS entries (DeepSeek, Qwen, Llama, Mistral, GLM) go through — the OpenAI-shaped
-// endpoint 404s/"malformed argument"s for Claude models on at least some projects.
-function isClaudeModel(model: string) {
-  return model.toLowerCase().startsWith("claude-");
-}
-
-// Defensive normalizer: target-format resolution for manually-added custom Claude models under
-// "vertex"/"vertex-partner" was observed sending a Gemini-shaped body (contents/parts) to the
-// Anthropic rawPredict endpoint instead of the configured "claude" format, causing a hard
-// "messages: Field required" error upstream regardless of the stored per-model targetFormat. This
-// converts a Gemini-shaped body to Anthropic Messages shape so the executor works either way,
-// independent of that unresolved upstream resolution gap.
-function toAnthropicBody(body: Record<string, unknown>): Record<string, unknown> {
-  const contents = body.contents as Array<{ role?: string; parts?: Array<{ text?: string }> }> | undefined;
-  if (!Array.isArray(contents)) return body;
-
-  const messages = contents.map((c) => ({
-    role: c.role === "model" ? "assistant" : "user",
-    content: (c.parts || []).map((p) => p.text || "").join(""),
-  }));
-  const generationConfig = body.generationConfig as { maxOutputTokens?: number } | undefined;
-  const systemInstruction = body.systemInstruction as { parts?: Array<{ text?: string }> } | undefined;
-
-  const converted: Record<string, unknown> = {
-    messages,
-    max_tokens: generationConfig?.maxOutputTokens || 4096,
-  };
-  if (systemInstruction?.parts?.length) {
-    converted.system = systemInstruction.parts.map((p) => p.text || "").join("");
-  }
-  return converted;
-}
-
-// rawPredict always returns a single complete JSON body, never real SSE framing (see buildUrl).
-// When the caller actually requested a stream, synthesize a genuine Anthropic-native event
-// sequence from that JSON so the existing claude-to-openai.ts (and sibling) response translators
-// — which already parse real message_start/content_block_*/message_delta/message_stop events —
-// can consume it correctly, instead of relying on the OpenAI-`choices`-only JSON→SSE fallback
-// (open-sse/utils/jsonToSse.ts) which cannot represent Anthropic's native response shape at all.
-function synthesizeClaudeSse(response: Record<string, unknown>): string {
-  const messageId = typeof response.id === "string" ? response.id : `msg_${Date.now()}`;
-  const model = typeof response.model === "string" ? response.model : "";
-  const usage = (response.usage as Record<string, unknown>) || {};
-  const stopReason = typeof response.stop_reason === "string" ? response.stop_reason : "end_turn";
-  const stopSequence = (response.stop_sequence as string | null | undefined) ?? null;
-  const content = Array.isArray(response.content) ? response.content : [];
-
-  const events: Array<{ event: string; data: Record<string, unknown> }> = [];
-
-  events.push({
-    event: "message_start",
-    data: {
-      type: "message_start",
-      message: {
-        id: messageId,
-        type: "message",
-        role: "assistant",
-        content: [],
-        model,
-        stop_reason: null,
-        stop_sequence: null,
-        usage: { input_tokens: usage.input_tokens || 0, output_tokens: 0 },
-      },
-    },
-  });
-
-  content.forEach((block: Record<string, unknown>, index: number) => {
-    if (block.type === "text") {
-      events.push({
-        event: "content_block_start",
-        data: { type: "content_block_start", index, content_block: { type: "text", text: "" } },
-      });
-      if (block.text) {
-        events.push({
-          event: "content_block_delta",
-          data: {
-            type: "content_block_delta",
-            index,
-            delta: { type: "text_delta", text: block.text },
-          },
-        });
-      }
-      events.push({ event: "content_block_stop", data: { type: "content_block_stop", index } });
-    } else if (block.type === "tool_use") {
-      events.push({
-        event: "content_block_start",
-        data: {
-          type: "content_block_start",
-          index,
-          content_block: { type: "tool_use", id: block.id, name: block.name, input: {} },
-        },
-      });
-      events.push({
-        event: "content_block_delta",
-        data: {
-          type: "content_block_delta",
-          index,
-          delta: { type: "input_json_delta", partial_json: JSON.stringify(block.input ?? {}) },
-        },
-      });
-      events.push({ event: "content_block_stop", data: { type: "content_block_stop", index } });
-    } else if (block.type === "thinking") {
-      events.push({
-        event: "content_block_start",
-        data: { type: "content_block_start", index, content_block: { type: "thinking", thinking: "" } },
-      });
-      if (block.thinking) {
-        events.push({
-          event: "content_block_delta",
-          data: {
-            type: "content_block_delta",
-            index,
-            delta: { type: "thinking_delta", thinking: block.thinking },
-          },
-        });
-      }
-      events.push({ event: "content_block_stop", data: { type: "content_block_stop", index } });
-    }
-  });
-
-  events.push({
-    event: "message_delta",
-    data: {
-      type: "message_delta",
-      delta: { stop_reason: stopReason, stop_sequence: stopSequence },
-      usage: { output_tokens: usage.output_tokens || 0 },
-    },
-  });
-
-  events.push({ event: "message_stop", data: { type: "message_stop" } });
-
-  return events.map((e) => `event: ${e.event}\ndata: ${JSON.stringify(e.data)}\n\n`).join("");
-}
-
 export class VertexExecutor extends BaseExecutor {
   constructor() {
     super("vertex", PROVIDERS.vertex);
   }
 
   async execute(input: ExecuteInput) {
-    const { credentials, log, model, stream } = input;
+    const { credentials, log } = input;
     // Defensive: trim stray surrounding whitespace from a pasted credential.
     if (typeof credentials.apiKey === "string") {
       credentials.apiKey = credentials.apiKey.trim();
@@ -296,53 +160,7 @@ export class VertexExecutor extends BaseExecutor {
         throw err;
       }
     }
-    if (isClaudeModel(model) && input.body && typeof input.body === "object") {
-      let body = input.body as Record<string, unknown>;
-      if (!Array.isArray(body.messages)) {
-        body = toAnthropicBody(body);
-        input.body = body;
-      }
-      // The rawPredict endpoint requires "anthropic_version" in the body (Vertex's substitute
-      // for the "anthropic-version" header used by Anthropic's direct API).
-      body.anthropic_version ??= "vertex-2023-10-16";
-      // Unlike Anthropic's direct API (which reads the model from the body), Vertex's
-      // rawPredict endpoint already encodes project/region/model in the URL and 400s with
-      // "model: Extra inputs are not permitted" if the translated request body still carries
-      // one (the openai→claude request translator copies the client's model field over).
-      delete body.model;
-    }
-
-    const result = await super.execute(input);
-
-    if (isClaudeModel(model) && stream) {
-      const response = result instanceof Response ? result : result?.response;
-      if (response?.ok) {
-        const contentType = response.headers.get("content-type") || "";
-        if (contentType.includes("application/json") && !contentType.includes("text/event-stream")) {
-          const jsonText = await response.text();
-          let newBody = jsonText;
-          let newContentType = contentType;
-          try {
-            newBody = synthesizeClaudeSse(JSON.parse(jsonText));
-            newContentType = "text/event-stream";
-          } catch (err) {
-            const message = err instanceof Error ? err.message : String(err);
-            log?.warn?.("VERTEX", `Failed to synthesize Claude SSE stream: ${message}`);
-          }
-          const newHeaders = new Headers(response.headers);
-          newHeaders.set("content-type", newContentType);
-          newHeaders.delete("content-length");
-          const newResponse = new Response(newBody, {
-            status: response.status,
-            statusText: response.statusText,
-            headers: newHeaders,
-          });
-          return result instanceof Response ? newResponse : { ...result, response: newResponse };
-        }
-      }
-    }
-
-    return result;
+    return super.execute(input);
   }
 
   buildUrl(model: string, stream: boolean, urlIndex = 0, credentials: any = null) {
@@ -371,13 +189,6 @@ export class VertexExecutor extends BaseExecutor {
       }
     }
 
-    if (isClaudeModel(model)) {
-      // streamRawPredict?alt=sse was verified to return a single plain JSON body (not real SSE
-      // framing) rather than actual chunked events, which breaks the SSE parser upstream
-      // ("stream ended before producing a non-ping SSE event"). rawPredict is confirmed reliable
-      // for both streaming and non-streaming requests; always use it here.
-      return `https://aiplatform.googleapis.com/v1/projects/${project}/locations/${region}/publishers/anthropic/models/${model}:rawPredict`;
-    }
     if (isPartnerModel(model)) {
       return `https://aiplatform.googleapis.com/v1/projects/${project}/locations/global/endpoints/openapi/chat/completions`;
     }

open-sse/executors/yuanbao-web.ts

@@ -457,12 +457,9 @@ function transformYuanbaoStream(
             if (event.type === "think" && event.content) {
               ensureRole();
               emit({ reasoning_content: event.content });
-            } else if (event.type === "text") {
-              const text = event.msg ?? event.content;
-              if (text) {
-                ensureRole();
-                emit({ content: text });
-              }
+            } else if (event.type === "text" && typeof event.msg === "string" && event.msg) {
+              ensureRole();
+              emit({ content: event.msg });
             }
           }
         }
@@ -501,10 +498,7 @@ async function collectYuanbaoResponse(
         const event = parseYuanbaoDataLine(line);
         if (!event) continue;
         if (event.type === "think" && event.content) reasoning += event.content;
-        else if (event.type === "text") {
-          const text = event.msg ?? event.content;
-          if (text) content += text;
-        }
+        else if (event.type === "text" && typeof event.msg === "string") content += event.msg;
       }
     }
   } finally {

open-sse/handlers/audioSpeech.ts

@@ -228,35 +228,6 @@ async function handleDeepgramSpeech(providerConfig, body, modelId, token) {
   return audioStreamResponse(res);
 }
 
-/**
- * Handle Soniox TTS (OpenAI speech shape → Soniox /tts, returns raw audio bytes)
- */
-async function handleSonioxSpeech(providerConfig, body, modelId, token) {
-  const fmt = typeof body.response_format === "string" ? body.response_format : "mp3";
-  const audioFormat = fmt === "pcm" ? "pcm_s16le" : fmt;
-
-  const res = await fetch(providerConfig.baseUrl, {
-    method: "POST",
-    headers: {
-      "Content-Type": "application/json",
-      ...buildAuthHeaders(providerConfig, token),
-    },
-    body: JSON.stringify({
-      text: body.input,
-      model: modelId,
-      ...(body.voice ? { voice: body.voice } : {}),
-      audio_format: audioFormat,
-    }),
-  });
-
-  if (!res.ok) {
-    return upstreamErrorResponse(res, await res.text());
-  }
-
-  const contentType = fmt === "wav" ? "audio/wav" : fmt === "opus" ? "audio/opus" : "audio/mpeg";
-  return audioStreamResponse(res, contentType);
-}
-
 /**
  * Handle ElevenLabs TTS
  * POST {baseUrl}/{voice_id} with { text, model_id }
@@ -875,10 +846,6 @@ export async function handleAudioSpeech({
       return handleDeepgramSpeech(providerConfig, body, modelId, token);
     }
 
-    if (providerConfig.format === "soniox-tts") {
-      return handleSonioxSpeech(providerConfig, body, modelId, token);
-    }
-
     if (providerConfig.format === "elevenlabs") {
       return handleElevenLabsSpeech(providerConfig, body, modelId, token);
     }

open-sse/handlers/audioTranscription.ts

@@ -336,78 +336,6 @@ async function handleGladiaTranscription(providerConfig, file, modelId, token) {
   return errorResponse(504, "Gladia transcription timed out after 120s");
 }
 
-/**
- * Handle Soniox transcription (async: upload file → create job → poll → get transcript)
- */
-async function handleSonioxTranscription(providerConfig, file, modelId, token) {
-  const authHeaders = buildAuthHeaders(providerConfig, token);
-
-  const { body: uploadBody, contentType: uploadContentType } = await buildMultipartBody(file, {});
-  const uploadRes = await fetch("https://api.soniox.com/v1/files", {
-    method: "POST",
-    headers: { ...authHeaders, "Content-Type": uploadContentType },
-    body: uploadBody,
-  });
-  if (!uploadRes.ok) {
-    return upstreamErrorResponse(uploadRes, await uploadRes.text());
-  }
-  const fileId = (await uploadRes.json()).id;
-
-  const createRes = await fetch(providerConfig.baseUrl, {
-    method: "POST",
-    headers: { ...authHeaders, "Content-Type": "application/json" },
-    body: JSON.stringify({
-      model: modelId,
-      file_id: fileId,
-      enable_language_identification: true,
-    }),
-  });
-  if (!createRes.ok) {
-    return upstreamErrorResponse(createRes, await createRes.text());
-  }
-  const { id: transcriptionId } = await createRes.json();
-
-  const statusUrl = `${providerConfig.baseUrl}/${transcriptionId}`;
-  const maxWait = 120_000;
-  const start = Date.now();
-  let completed = false;
-  while (Date.now() - start < maxWait) {
-    await new Promise((r) => setTimeout(r, 2000));
-    const pollRes = await fetch(statusUrl, { headers: authHeaders });
-    if (!pollRes.ok) {
-      continue;
-    }
-    const result = await pollRes.json();
-    if (result.status === "completed") {
-      completed = true;
-      break;
-    }
-    if (result.status === "error") {
-      return errorResponse(
-        500,
-        result.error_message || result.error || "Soniox transcription failed"
-      );
-    }
-  }
-  if (!completed) {
-    return errorResponse(504, "Soniox transcription timed out after 120s");
-  }
-
-  const transcriptRes = await fetch(`${statusUrl}/transcript`, { headers: authHeaders });
-  if (!transcriptRes.ok) {
-    return upstreamErrorResponse(transcriptRes, await transcriptRes.text());
-  }
-  const transcript = await transcriptRes.json();
-  const text =
-    typeof transcript.text === "string" && transcript.text.length > 0
-      ? transcript.text
-      : Array.isArray(transcript.tokens)
-        ? transcript.tokens.map((t: { text?: string }) => t.text ?? "").join("")
-        : "";
-
-  return Response.json({ text }, { headers: { ...CORS_HEADERS } });
-}
-
 /**
  * Handle Nvidia NIM transcription
  * Multipart POST, transform response to { text }
@@ -807,10 +735,6 @@ export async function handleAudioTranscription({
     return handleGladiaTranscription(providerConfig, file, modelId, token);
   }
 
-  if (providerConfig.format === "soniox") {
-    return handleSonioxTranscription(providerConfig, file, modelId, token);
-  }
-
   if (providerConfig.format === "nvidia-asr") {
     return handleNvidiaTranscription(providerConfig, file, modelId, token);
   }

open-sse/handlers/chatCore.ts

@@ -490,10 +490,9 @@ export async function handleChatCore({
     model,
     provider,
     apiKeyInfo,
-    headers: clientRawRequest?.headers,
     log,
   });
-  if (pluginGate.blocked === true) {
+  if (pluginGate.blocked) {
     return {
       success: false,
       status: 403,
@@ -1884,7 +1883,7 @@ export async function handleChatCore({
     modelOutputCap,
     toPositiveInteger(resolveInputTokenCapForGate({ provider, model: effectiveModel }, { isCombo }))
   );
-  if (outputBudget.ok === false) {
+  if (!outputBudget.ok) {
     const exceededInputCap = outputBudget.maxInputTokens !== undefined;
     const message =
       `Input exceeds ${exceededInputCap ? "maximum input tokens" : "context window"} for ${provider}/${effectiveModel}: ` +
@@ -4623,7 +4622,6 @@ export async function handleChatCore({
       model,
       provider,
       apiKeyInfo,
-      headers: clientRawRequest?.headers,
       response: { status: 200, data: translatedResponse },
     });
 
@@ -5014,7 +5012,6 @@ export async function handleChatCore({
     model,
     provider,
     apiKeyInfo,
-    headers: clientRawRequest?.headers,
     response: { status: 200, streamed: true },
   });
 

open-sse/handlers/chatCore/pluginOnRequest.ts

@@ -10,10 +10,13 @@
  */
 
 type LoggerLike =
-  { info?: (...args: unknown[]) => void; debug?: (...args: unknown[]) => void } | null | undefined;
+  | { info?: (...args: unknown[]) => void; debug?: (...args: unknown[]) => void }
+  | null
+  | undefined;
 
 export type PluginOnRequestGate =
-  { blocked: true; response: Response } | { blocked: false; body?: unknown };
+  | { blocked: true; response: Response }
+  | { blocked: false; body?: unknown };
 
 const JSON_HEADERS = { status: 403, headers: { "Content-Type": "application/json" } } as const;
 
@@ -23,7 +26,6 @@ export async function runPluginOnRequestHook(args: {
   model: string | null | undefined;
   provider: string | null | undefined;
   apiKeyInfo: unknown;
-  headers?: Record<string, string | string[] | undefined>;
   log?: LoggerLike;
 }): Promise<PluginOnRequestGate> {
   try {
@@ -34,7 +36,6 @@ export async function runPluginOnRequestHook(args: {
       model: args.model,
       provider: args.provider,
       apiKeyInfo: args.apiKeyInfo,
-      headers: args.headers,
       metadata: {},
     };
     const pluginResult = await runOnRequest(pluginCtx);

open-sse/handlers/chatCore/pluginOnResponse.ts

@@ -24,7 +24,6 @@ export async function runPluginOnResponseHook(args: {
   model: string | null | undefined;
   provider: string | null | undefined;
   apiKeyInfo: unknown;
-  headers?: Record<string, string | string[] | undefined>;
   response: PluginOnResponsePayload;
 }): Promise<void> {
   try {
@@ -36,7 +35,6 @@ export async function runPluginOnResponseHook(args: {
         model: args.model,
         provider: args.provider,
         apiKeyInfo: args.apiKeyInfo,
-        headers: args.headers,
         metadata: {},
       },
       args.response

open-sse/handlers/chatCore/sanitization.ts

@@ -46,7 +46,7 @@ export function sanitizeChatRequestBody(
   }
 
   if (Array.isArray(body.tools)) {
-    const tools = body.tools.filter((tool: Record<string, unknown>) => {
+    body.tools = body.tools.filter((tool: Record<string, unknown>) => {
       const toolType = typeof tool.type === "string" ? tool.type : "";
       if (toolType && toolType !== "function" && !tool.function && tool.name === undefined) {
         return true;
@@ -56,7 +56,7 @@ export function sanitizeChatRequestBody(
       return name && String(name).trim().length > 0;
     });
 
-    body.tools = tools.map((tool) => sanitizeOpenAITool(tool));
+    body.tools = body.tools.map((tool) => sanitizeOpenAITool(tool) as (typeof body.tools)[number]);
   }
 
   return body;

open-sse/handlers/chatCore/targetFormat.ts

@@ -46,15 +46,13 @@ export function resolveChatCoreTargetFormat(opts: {
       sourceFormat === FORMATS.CLAUDE)
       ? sourceFormat
       : undefined;
-  // #8994: model-level targetFormat overrides (from registry or custom-model DB override)
-  // take precedence over apiFormat="responses" — otherwise Vertex Claude models with
-  // targetFormat="claude" get wrongly routed to OpenAI Responses format.
   let targetFormat =
-    modelTargetFormat ||
-    customModelTargetFormat ||
-    (apiFormat === "responses"
+    apiFormat === "responses"
       ? FORMATS.OPENAI_RESPONSES
-      : inferredAgentRouterTargetFormat || getTargetFormat(provider, providerSpecificData));
+      : modelTargetFormat ||
+        customModelTargetFormat ||
+        inferredAgentRouterTargetFormat ||
+        getTargetFormat(provider, providerSpecificData);
   if (nativeXaiResponsesPassthrough) targetFormat = FORMATS.OPENAI_RESPONSES;
   return { alias, targetFormat };
 }