Claude-skill-registry idea-generation
Work with IdeaForge's AI idea generation system. Triggers: generation flow, AI prompts, scoring system, duplicate detection, real-time logs, generation debugging. Pipeline: API → PromptBuilder → AI → Parse → Dedupe → Save.
install
source · Clone the upstream repo
git clone https://github.com/majiayu000/claude-skill-registry
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/majiayu000/claude-skill-registry "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/data/idea-generation" ~/.claude/skills/majiayu000-claude-skill-registry-idea-generation && rm -rf "$T"
manifest:
skills/data/idea-generation/SKILL.mdtags
source content
Idea Generation
Pipeline
POST /api/generation/generate → IdeaGenerationService → PromptBuilder (YAML configs) → callAI() (Claude/Gemini) → parseAIResponse() → EmbeddingService (duplicate check) → IdeaRepository (save) → GenerationLogger (SSE logs)
API
POST /api/generation/generate { "framework": "pain-point", // optional, random if not specified "domain": "Healthcare", // optional "sessionId": "uuid" // for log tracking }
AI Response Shape
{ "name": "Idea Name (max 60 chars)", "domain": "Domain → Subdomain", "problem": "...", "solution": "...", "quickSummary": "Elevator pitch", "concreteExample": { "currentState": "How users handle this today", "yourSolution": "How they'd use your product", "keyImprovement": "Quantifiable improvement" }, "evaluation": { "problemSeverity": { "score": 8, "reasoning": "..." } }, "tags": ["tag1", "tag2"] }
Scoring
- Per-criterion: 1-10 scale
- Weighted total: 0-100 scale
- Config:
evaluation-criteria.yaml
Duplicate Detection
- Exact: Same domain + problem + solution
- Semantic: Embedding similarity > 85%
SSE Logs
Stages:
INIT → CONFIG_LOAD → PROMPT_BUILD → API_CALL → RESPONSE_PARSE → DUPLICATE_CHECK → DB_SAVE → COMPLETE
Endpoint:
GET /api/logs/stream/{sessionId}
Key Files
| File | Purpose |
|---|---|
| Orchestration |
| AI prompts |
| YAML configs |
| Claude/Gemini |