Skip to content

AGENTS.md

Latest commit

 

History

History
29 lines (24 loc) · 1.67 KB

AGENTS.md

File metadata and controls

29 lines (24 loc) · 1.67 KB

AGENTS.md

Purpose

Codex-first operating rules for this repository. This file replaces Claude hook-based auto-activation with explicit trigger rules and repeatable workflows.

Skill Trigger Rules

  • Use skills/backend-dev-guidelines when task includes API routes, controllers, services, repositories, Prisma, middleware, or backend validation.
  • Use skills/frontend-dev-guidelines when task includes React components, MUI, routing, data fetching, frontend state, or UI performance.
  • Use skills/route-tester when task includes authenticated endpoint testing, cookie/JWT auth, or route verification.
  • Use skills/error-tracking when task includes Sentry setup, exception capture, monitoring, or production-debug context.
  • Use skills/skill-developer when creating/updating skills, trigger rules, or capability packaging.

Guardrail Rules (Codex)

  • Before frontend code edits, verify MUI version compatibility and avoid legacy v5/v6 API patterns.
  • Before major backend edits, enforce layered boundaries: route -> controller -> service -> repository.
  • Before finishing a task, run available project checks (at minimum typecheck and build if defined).
  • 코드 주석은 기본적으로 한글로 작성한다.

Working Protocol

  1. Identify applicable skill(s) from trigger rules.
  2. Read only required skill sections/resources (progressive disclosure).
  3. Apply changes.
  4. Run validation commands.
  5. Record outcomes and next actions in dev/active/<task>/ when work is non-trivial.

Templates

  • Planning: templates/dev-docs.template.md
  • Context update: templates/dev-docs-update.template.md
  • Route testing research: templates/route-research-for-testing.template.md