GitLawB OpenGateway API (xiaomi-mimo compatible) does not expose a /models
endpoint, causing validateOpenAILikeProvider to 404 on the initial probe
and report 'Provider validation endpoint not supported'.
Add specialty validators for both gitlawb and gitlawb-gmi that follow the
same pattern as the existing xiaomi-mimo validator: skip GET /models,
validate directly via POST /chat/completions with a minimal test message.
Any 401/403 response means an invalid key; all other responses mean auth
is OK.
Fixes test-connection returning 404 for GitLawB providers.
Rule #16 previously banned all `Co-Authored-By` trailers absolutely.
That blocked the upstream-port workflows (`/port-upstream-features` and
`/port-upstream-issues`), which must credit human upstream PR authors
and issue reporters in OmniRoute commits.
Refine the rule to ban only AI/bot-attributed trailers (Claude, GPT,
Copilot, Bot; anthropic.com / openai.com / bot-owned noreply.github.com
emails) while allowing standard human `Co-authored-by: Name <email>`
attribution.
Sync the rule across the source CLAUDE.md, the E2E shakedown doc note,
and 41 i18n translations.
Base release/v3.8.6 already below the historical 75/75/75/70 threshold
(67.6% statements pre-existing per F1 audit). Group C adds 100%-covered
new code locally; the relaxed gate lets the release land without
masking that pre-existing debt. Aspirational gate of 75/75/75/70 stays
in CLAUDE.md; restoration is a follow-up initiative.
Temporarily relaxes the c8 thresholds and Hard Rule #9 from
75/75/75/70 to 40/40/40/40 across statements/lines/functions/branches
so the v3.8.6 page-redesign branches (translator, playground, search-tools,
batch, memory, monitoring) can merge before reaching their final test
coverage targets. Update upward as the new pages mature.
Owner-authorized temporary relaxation to unblock landing of planos 16+22
(Monitoring reorg + Quota Share Engine). Restore to 75/75/75/70 after
Group B catches up coverage in follow-up PR. Critical modules
(fairShare, sqliteQuotaStore, enforce, ...) keep local target >=90%.
- RiskNoticeBanner: use lazy useState initializer for localStorage read
instead of useEffect + setState
- ModelSelectorModal: move fetch logic to useCallback, call from useEffect
to avoid setState directly in effect body
AgentSkillsPageClient (F7) was rewritten and no longer imports AGENT_SKILLS.
Grep confirms zero remaining imports of the symbol in src/, open-sse/, tests/,
or electron/. Remove the backward-compatible shim and TODO(F7) comment.
Replace the stale 18-entry omniroute-* index (with broken links to pruned
skill directories) with a complete table of all 42 current skills (22 API +
20 CLI). Adds entry-point pointers, raw URL pattern, agent-discovery section
(MCP tool + A2A skill), and cross-reference to docs/frameworks/AGENT-SKILLS.md.
The /api/agent-skills/[id]/raw endpoint returns text/markdown (plain string),
not a JSON envelope. The client was incorrectly calling res.json() and
unpacking a non-existent `body` field, causing all preview panes to show
empty content. Switch to res.text() and update the test mock to match.
