Agent skill
treido-orchestrator
Orchestrates Treido work end-to-end (parallel specialist audits, single-writer plan, execute, verify/test). Uses a bundle matrix to spawn auditors. Trigger: ORCH
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/data/treido-orchestrator
SKILL.md
Treido Orchestrator (Single Writer)
This skill coordinates a predictable workflow:
- Phase 1 (Parallel): specialist auditors (read-only) return structured payloads
- Phase 2 (Single-writer): merge audits + write a prioritized plan in
TASKS.md - Phase 3 (Single-writer): implement tasks in 1–3 file batches
- Phase 4 (Single-writer): run verification gates
- Phase 5 (Single-writer): run the smallest relevant tests
Non-Negotiables
- Single-writer rule:
- Auditors + verifier are read-only (no patches, no
TASKS.mdedits). - Orchestrator is the sole writer for
audit/*andTASKS.md. - Code changes happen in the implementer lane (still one writer per batch).
- Auditors + verifier are read-only (no patches, no
- Small batches: execute as 1–3 file rounds; run gates after each round.
- Rails first: no secrets/PII in logs, next-intl for user strings, Tailwind v4 tokens only, cached-server rules, no scope creep.
Bundle Matrix (Intent → Auditors)
| User intent (examples) | Bundle | Auditors spawned |
|---|---|---|
| “fix styling”, “UI looks off”, “mobile styling” | UI | Next.js + TW4 + SHADCN |
| “backend issue”, “server action broken”, “webhook” | Backend | Next.js + Supabase + TypeScript |
| “database”, “RLS”, “schema/policies” | Supabase | Supabase + TypeScript |
| “full audit”, “what’s wrong” | Full | All: Next.js + TW4 + SHADCN + Supabase + TypeScript |
Preferred auditor skill names (new system):
treido-audit-nextjstreido-audit-tw4treido-audit-shadcntreido-audit-supabasetreido-audit-typescript
Fallbacks (if a new auditor skill doesn’t exist yet):
- Next.js:
nextjs-structure(ortreido-frontendfor boundary scans) - TW4:
treido-tailwind-v4 - SHADCN:
treido-shadcn - Supabase:
treido-supabase-mcp - TypeScript:
treido-audit(TS-focused scan) +pnpm -s typecheck
Phase 1 — Specialist Audits (Parallel, Read-only)
- Select a bundle from the matrix (or accept an explicit override like
ORCH:UI,ORCH:FULL). - Spawn auditors in parallel. Each auditor must return only the structured payload contract:
- Contract:
.codex/skills/treido-orchestrator/references/audit-payload.md
- Contract:
- Validate the payloads are mergeable (each starts with
## <AUDITOR>and includes the required sections).
Auditor Prompt Template (copy/paste)
Send each auditor this message (customize scope/context):
You are an audit-only specialist. Read-only: do not patch files, do not edit TASKS.md.
Goal/context: <1–2 sentences>
Scope hints: <routes/files if known>
Return ONLY a Markdown section that follows:
.codex/skills/treido-orchestrator/references/audit-payload.md
No preamble. Start with `## <AUDITOR_NAME>`.
Phase 1.5 — Merge Audits (Single Writer)
- Create a shared audit artifact under
audit/:audit/YYYY-MM-DD_<short-context>.md
- Merge auditor sections verbatim (no rewrites beyond formatting), ordered by bundle.
- Ensure no secrets/PII are written into audit artifacts.
Phase 2 — Plan (Write TASKS.md)
- Convert findings into tasks with clear priority and acceptance criteria.
- Keep tasks small (≤ 1 day) and executable in 1–3 file rounds.
- Prefer these owners (new system), otherwise use existing owners:
treido-impl-frontend(fallback:treido-frontend)treido-impl-backend(fallback:treido-backend)treido-verify(fallback:treido-audit)
- Respect
TASKS.md“Ready ≤15” rules.
Task block (recommended):
### Priority 1 (Critical)
- [ ] <ID>: <task summary>
- Owner: <skill>
- Verify: `<commands>`
- Files: <paths>
Phase 3 — Execute (Single Writer)
- Pick the smallest set of Priority 1 tasks that fit into a 1–3 file batch.
- Implement using the owner lane conventions (frontend vs backend).
- Avoid unrelated refactors; defer follow-ups into
TASKS.md.
Phase 4 — Verify / QA
Always:
pnpm -s typecheck
pnpm -s lint
pnpm -s styles:gate
Phase 5 — Tests (Risk-based)
- UI-only: consider
pnpm -s test:unit(and/or Storybook spot-check if used) - Auth/payments/checkout/webhooks:
REUSE_EXISTING_SERVER=true pnpm -s test:e2e:smoke - Database/RLS: run Supabase advisors + regenerate types (via MCP lane) when applicable
References (load only if needed)
- SSOT:
AGENTS.md,docs/ARCHITECTURE.md,docs/DESIGN.md - Audit payload contract:
.codex/skills/treido-orchestrator/references/audit-payload.md - Existing lanes (fallbacks):
.codex/skills/treido-frontend/SKILL.md,.codex/skills/treido-backend/SKILL.md,.codex/skills/treido-supabase-mcp/SKILL.md,.codex/skills/treido-audit/SKILL.md
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
Didn't find tool you were looking for?