Implements the catalog.ts public API defined in §3.3 of the master plan:
- getCatalog(): AgentSkill[] — returns 42 entries, lazy-cached in module scope
- getSkillById(id): AgentSkill | null — lookup by canonical ID
- filterCatalog(opts): AgentSkill[] — filter by category and/or area
- computeCoverage(): SkillCoverage — reads skills/ dir and counts SKILL.md present
- refreshCatalog(): void — invalidates cache (used by tests + generator)
- fetchSkillMarkdown(id): Promise<SkillMarkdown> — reads local fs first,
falls back to GitHub raw fetch with 1h Next.js cache (for F4 /raw route)
API_SKILL_IDS and CLI_SKILL_IDS exported as readonly string arrays (D28 order).
Single source of truth for all consumers (REST routes, MCP, A2A).
Replace 18-entry hardcoded AGENT_SKILLS array with 42-entry CURATED_SKILLS
covering all areas listed in D28 of the master plan. Each curated entry
has id, name, description, category, area, icon, and optional flags.
Adds backward-compatible AGENT_SKILLS alias (deprecated) that maps curated
entries to the old AgentSkill shape with empty endpoints/cliCommands arrays
so the existing /dashboard/agent-skills page continues to work until F7
rewrites it.
Imports AgentSkill, SkillArea, SkillCategory types from src/lib/agentSkills/types.ts
(F1 single source of truth) instead of redeclaring locally.
Add qwen3.7-max, mimo-v2-pro, mimo-v2-omni, hy3-preview to the
opencode-go provider. Extend executor tests and model specs.
- Registry now matches live https://opencode.ai/zen/go/v1/models response
(16 models). Previously, four models returned by the upstream API were
absent from the static registry, so requests to them failed model
resolution before reaching the executor.
- qwen3.7-max: inherits defaultContextLength 200000; model spec mirrors
qwen3.6-plus (Bailian multimodal, thinking + tools + vision).
- mimo-v2-pro: spec mirrors mimo-v2-omni (262144 ctx, 131072 maxOut,
tools + vision).
- mimo-v2-omni: registry entry only (spec already present upstream).
- hy3-preview: registry entry only; falls back to __default__ spec.
- All four route to /chat/completions (default openai format), matching
the existing opencode-go executor behavior.
Replace __MISSING__ placeholders with Simplified Chinese translations
across webhooks (wizard/howItWorks/deliveries), costs, endpoint, health,
logs, providers, settings, usage, sidebar and related modules.
- Fix 28 broken flat doc path links in README.md to use nested subdirectory paths (guides/, reference/, ops/, compression/, architecture/, routing/, frameworks/)
- Update stale project-structure tree in CONTRIBUTING.md to reflect current nested docs/ layout
- Remove duplicate docs/AUTO-COMBO.md (already exists at docs/routing/AUTO-COMBO.md)
