Find, evaluate, and claim open-source bounties in seconds. HeadHunter scans bounty platforms, scores opportunities by reward-to-effort ratio, and generates solution blueprints so you can start coding immediately.
Enter your GitHub username and see your skill profile matched against live bounties.
HeadHunter is an AI bounty hunting agent that eliminates the tedious process of discovering profitable open-source bounties. Instead of manually scrolling bounty boards, clicking into repos, and evaluating whether a bounty is worth your time, HeadHunter does the triage for you. One API call returns a ranked list of bounties you can actually claim and profit from.
Built for the Purch Vault Hackathon (Mar 20 - Apr 20, 2026). Submitting to: Best Purch Integration + Best Skill.
- Multi-platform scanning: Crawls GitHub Issues and Algora for open bounties in real time
- Smart scoring: Ranks bounties by reward/effort ratio, skill match, competition level, recency, and maintainer activity
- Deep evaluation: Analyzes repos, checks competing PRs, estimates effort, returns GO/CAUTION/SKIP verdict
- Solution blueprints: Generates step-by-step implementation plans with key files, LOC estimates, and PR templates
- Developer profiling: Extracts skill profiles from GitHub activity to match bounties to your strengths
- x402 micropayments: Pay-per-query via USDC on Base. Other AI agents can call HeadHunter autonomously
- Claude Code skill: Install as
/headhunterin Claude Code for bounty hunting from the terminal
Developer HeadHunter API Bounty Platforms
| | |
|-- /profile ------------>| |
|<-- skill extraction ----| |
| | |
|-- /hunt --------------->|-- scan ------------------>|
| |<-- bounties --------------|
|<-- ranked results ------| |
| | |
|-- /evaluate ----------->|-- repo analysis --------->|
|<-- GO/CAUTION/SKIP -----| |
| | |
|-- /strategize --------->|-- codebase reading ------>|
|<-- implementation plan -| |
| | |
|-- /solve (via skill) -->| (fork, implement, PR) |
| Method | Endpoint | Price | Description |
|---|---|---|---|
| POST | /hunt |
$0.02 | Find bounties matching your skills |
| POST | /evaluate |
$0.08 | Deep-analyze a specific bounty |
| POST | /strategize |
$0.15 | Generate a solution blueprint |
| POST | /profile |
Free | Extract developer skill profile from GitHub |
curl -X POST https://headhunter-nine.vercel.app/hunt \
-H "Content-Type: application/json" \
-d '{"skills":["typescript","react"],"min_reward":100,"limit":5}'Response:
{
"bounties": [
{
"title": "Schema Migration System for ZIO Schema 2",
"platform": "algora",
"reward": "$4000",
"score": 8.8,
"competition": 0,
"url": "https://github.com/zio/zio-blocks/issues/519"
}
],
"total_found": 21,
"total_matched": 5,
"search_cost": "$0.02"
}Each bounty gets a composite score (0-10):
| Factor | Weight | What it measures |
|---|---|---|
| Reward/Effort | 25% | Dollar per hour. Higher pay = higher score |
| Skill Match | 20% | Overlap between your skills and bounty tags |
| Low Competition | 15% | Fewer competing PRs = better odds |
| Recency | 20% | Fresh bounties rank higher than stale ones |
| Maintainer Activity | 10% | Active maintainers review faster |
| Clarity | 10% | Clear specs with labels reduce risk |
HeadHunter is a paid x402 service. When an agent or user calls a paid endpoint:
- Client sends request to
/hunt,/evaluate, or/strategize - Server responds with
402 Payment Required+ x402 payment headers - Client wallet signs a USDC micropayment on Base
- Client resends request with payment proof
- Server verifies payment, executes query, returns results
This means other AI agents can autonomously pay HeadHunter to find bounties for their operators.
HeadHunter ships as an installable Claude Code skill:
# Install (symlink to your skills directory)
ln -s /path/to/headhunter/skill ~/.claude/skills/headhunterThen in Claude Code:
/headhunter hunt --skills typescript,react --min-reward 100
/headhunter evaluate https://github.com/org/repo/issues/123
/headhunter strategize https://github.com/org/repo/issues/123
/headhunter solve https://github.com/org/repo/issues/123
The /solve command runs the full pipeline: evaluate, fork, implement, and submit a PR to claim the bounty.
| Layer | Technology |
|---|---|
| Server | Hono (lightweight, edge-ready) |
| Language | TypeScript 5.8 |
| Runtime | Node.js |
| Payments | x402 protocol (USDC on Base) |
| GitHub API | Octokit |
| Bounty Sources | GitHub Issues, Algora |
| Skill System | Claude Code Skills |
git clone https://github.com/ajanaku1/headhunter.git
cd headhunter
npm installCreate .env from the example:
cp .env.example .envAdd your GitHub token (increases API rate limit from 60 to 5,000 req/hr):
GITHUB_TOKEN=ghp_your_token_here
Start the dev server:
npm run devServer runs at http://localhost:3402. Open in a browser to see the landing page, or call the API directly.
headhunter/
├── src/
│ ├── server.ts # Hono server with x402 middleware
│ ├── index.ts # Public API exports
│ ├── types.ts # TypeScript interfaces
│ ├── services/
│ │ ├── scanner.ts # Bounty hunt orchestrator
│ │ ├── evaluator.ts # Deep bounty analysis
│ │ ├── strategist.ts # Solution blueprint generator
│ │ └── profiler.ts # GitHub skill extraction
│ ├── platforms/
│ │ ├── github.ts # GitHub Issues scanner
│ │ └── algora.ts # Algora bounty scraper
│ ├── scoring/
│ │ └── scorer.ts # Composite scoring algorithm
│ └── utils/
│ ├── github-api.ts # Octokit client
│ └── cache.ts # In-memory response cache
├── skill/
│ └── SKILL.md # Claude Code skill manifest
├── public/
│ └── index.html # Landing page (Matrix theme)
├── .env.example
├── package.json
└── tsconfig.json
MIT