Files
digital-psychology/.ai/ai-contract.md
T
jackyu66gitandCursor dd94e57277 docs: freeze YuXinGu lexicon, journey, and P1 engineering sources
Establish product language contract, user journey with dual growth engines,
ERD/OpenAPI/Go boundaries, and H5 routes (/profile, /portrait, /relation,
/membership) before P1 implementation.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-02 16:12:10 +08:00

1017 B

AI Contract

Highest operational contract for every agent session.

Before coding

MUST read:

  1. .ai/constitution.md
  2. .ai/architecture.md
  3. .ai/domain.md (+ domain/domain-map.md when touching models/API)
  4. .ai/coding.md
  5. Task-relevant: api.md / database.md / ui.md / security.md / product/lexicon.md + product/feature-map.md
  6. UI tasks: design/design-system.md + design/component-catalog.md + design/platform/*
  7. Relevant ADR under .ai/adr/ if changing stack or style

After coding

MUST:

  1. Run / follow .ai/review.md and print Review block
  2. Verify .ai/definition-of-done.md
  3. Use matching checklist under .ai/checklists/
  4. Prefer verifying with .ai/commands.md (build/test/health)

Never

  • Guess API shape
  • Guess DB schema
  • Guess product requirement
  • Invent endpoints, tables, or domain words not in .ai/domain.md / OpenAPI / task
  • Silently reverse an Accepted ADR

If unclear

ASK FIRST. Do not invent. Do not “assume and continue”.