feat: AI Bounty Description Enhancer (solfoundry#848) [fj4WqyCCw3C5ShR1RfB7MoBPTpkRrBFYP1uT35g3MvT]

[waterWang]

Summary

Implements the AI Bounty Description Enhancer (Issue #848, 450K $FNDRY bounty).

Description

This feature adds an AI-powered system that analyzes vague bounty descriptions and generates improved versions with clearer requirements, acceptance criteria, and examples using multiple LLM providers.

Architecture

• New API endpoints under /api/bounties/{id}/enhance*:

  • POST /api/bounties/{id}/enhance — Trigger AI enhancement (Claude, Codex, Gemini)
  • GET /api/bounties/{id}/enhancements — List all enhancement records
  • GET /api/bounties/{id}/enhancements/{eid} — Get a specific enhancement
  • POST /api/bounties/{id}/enhancements/{eid}/approve — Approve & apply enhancement
  • POST /api/bounties/{id}/enhancements/{eid}/reject — Reject enhancement

• New service: app/services/bounty_enhancer_service.py — Core AI logic with:

  • Structured prompt engineering for each LLM provider
  • JSON response parsing with fallback handling
  • In-memory enhancement record storage
  • Approval workflow (approve updates bounty, reject marks as rejected)

• New models: app/models/bounty_enhance.py — Pydantic models for requests/responses

• Configuration: ANTHROPIC_API_KEY, OPENAI_API_KEY, GOOGLE_API_KEY env vars

Acceptance Criteria Met

1. ✅ Multi-LLM analysis (Claude via Anthropic, Codex via OpenAI, Gemini via Google)
2. ✅ Automatic description enhancement with structured sections
3. ✅ Maintainer approval workflow before publishing
4. ✅ 20 comprehensive tests (all passing)

Additional Fix

Fixed a pre-existing NameError: name List is not defined in app/api/bounties.py (missing List import from typing).

Testing

cd backend && python3 -m pytest tests/test_bounty_enhance.py -v
# 20 passed in 3.10s