Skip to content

Latest commit

 

History

History
41 lines (34 loc) · 1.89 KB

File metadata and controls

41 lines (34 loc) · 1.89 KB

Ralph Building Loop - Liquid Memory

You are running a Ralph BUILDING loop for: AIGC 创意工坊 - AI 提示词解析与管理工具

Goal

Implement a Next.js-based AI prompt analysis and management tool with the following capabilities:

  1. Image upload and AI-powered prompt reverse-engineering
  2. Local storage of creative items with thumbnail generation
  3. Structured prompt editor
  4. Tag system and search/filter
  5. Word network visualization (D3.js)
  6. SeeDream API integration for image generation

Context Files

  • agents/prd.json - User Stories with acceptance criteria (US-001 to US-014)
  • PRD.md - Full product requirements document
  • AGENTS.md - Build instructions, test commands, and operational learnings
  • IMPLEMENTATION_PLAN.md - Current implementation status and next tasks

Your Tasks Each Iteration

  1. Read IMPLEMENTATION_PLAN.md - Check current status and pick the next incomplete task
  2. Read agents/prd.json - Find the corresponding US (User Story) with acceptance criteria
  3. Investigate code - Don't assume files don't exist; check before creating
  4. Implement - Write/modify code to satisfy the acceptance criteria
  5. Run backpressure - Execute test/lint/typecheck commands from AGENTS.md
  6. Update IMPLEMENTATION_PLAN.md - Mark task complete, add notes
  7. Update AGENTS.md - Add any new operational learnings
  8. Commit - Use conventional commits format: feat: US-XXX description

Completion Criteria

  • All US-001 to US-014 marked as complete in IMPLEMENTATION_PLAN.md
  • TypeScript typecheck passes
  • All tests pass (if any)
  • Add line STATUS: COMPLETE to IMPLEMENTATION_PLAN.md when done

Important Notes

  • This is an open-source personal project - no commercial constraints
  • Use "Gremins" as the git commit author name
  • Prioritize working software over perfect architecture
  • Ask clarifying questions in IMPLEMENTATION_PLAN.md notes if blocked