🔭 Open Source Scout

Automated GitHub issue recommender for open source contributors. Scans a repo's issue tracker, scores issues by your focus areas and skill level, and uses AI to generate actionable recommendations.

How It Works

1. Fetches open issues from any GitHub repo via API
2. Scores each issue based on your configured focus areas, keywords, and labels
3. Ranks by relevance, freshness, engagement, and difficulty match
4. Generates a markdown report with AI-powered recommendations (optional)

Quick Start

# Clone
git clone https://github.com/yweiii/opensource-scout.git
cd opensource-scout

# Setup
pip install -r requirements.txt
cp config.example.yaml config.yaml
# Edit config.yaml with your focus areas

# Run
python scout.py                # With AI recommendations
python scout.py --no-ai       # Heuristic scoring only

Configuration

Edit config.yaml to set:

• repo — which GitHub repo to scan
• focus_areas — topics you care about, with labels, keywords, and weights
• tier — your current skill level (1=beginner → 4=maintainer)
• entry_labels — labels that indicate easy entry points
• top_n — how many issues to recommend

See config.example.yaml for a full example configured for Ray.

Automated Scanning (GitHub Actions)

The included GitHub Action runs every Monday and commits a fresh scout report to output/.

To enable AI-powered recommendations, add your Anthropic API key as a GitHub secret:

1. Go to repo Settings → Secrets and variables → Actions
2. Add ANTHROPIC_API_KEY with your key

You can also trigger a scan manually from the Actions tab → Weekly Scout → Run workflow.

Scoring Algorithm

Each issue is scored based on:

Signal	Impact
Focus area label match	+weight (configurable)
Focus area keyword match	+weight × 0.5
Entry-friendly label	+5
High engagement (5+ 👍)	+3
Recently updated (< 7 days)	+3
Already assigned	-5
Stale (90+ days)	-3

Example Output

🔭 Open Source Scout Report
Repo: ray-project/ray
Tier: 1
Focus: Ray Serve, LLM Serving, Performance

| # | Score | Difficulty | Title | Areas | 👍 | Updated |
|---|-------|-----------|-------|-------|-----|---------|
| #52746 | 18.0 | Medium | Ray Serve overhead for vLLM | Ray Serve, Performance | 12 | 2026-02-25 |
| #61125 | 15.0 | Easy | LLM batching config | LLM Serving | 3 | 2026-02-18 |

License

MIT