type: Knowledge Bundle title: PR Genius — Pre-submission PR Advisor description: Evidence-backed PR contribution advisor for large open-source projects version: 1.9.1 created: 2026-07-01 updated: 2026-09-05 author: zsxh1990 conforms_to: OKF v0.1 (Sudhakaran88/okf-conformance) + agent_guidelines extension

mcp-name: io.github.zsxh1990/pr-genius

PR Genius — The advisor that knows which PRs get closed

1355 loaded patterns across 61 repos. 100% quality pass rate. Clone → paste MCP config → ask "Should I open this PR to encode/httpx?"

🎯 What is PR Genius?

PR Genius is not a PR dashboard. It's an Outbound PR CRM for professional OSS contributors and AI agents:

Manage PRs you've submitted to other repos — when to fix CI, rebase, wait, ping, or abandon.

Status heartbeat runs daily via cron, auto-detecting:

• 🔴 NEEDS_REBASE / CI_FAILING — fix immediately
• 🟡 STALE_REVIEW — ping after threshold
• 🟡 STALE_NO_REVIEW — consider abandoning
• 🟢 CLEAN / WAITING — continue waiting

🛡️ Why PR Genius?

PR Genius doesn't write PRs for you. It knows which PRs get closed.

Real cases (PR Genius helped avoid these rejections):

🚀 Quick Start

pip install prgenius-core

# Analyze PR
python3 -m prgenius analyze "feat: add feature" --repo org/repo --body "Fixes #123"

# Coach (pass/fail)
python3 -m prgenius coach "feat: add feature" --repo org/repo

# Triage (policy check)
python3 -m prgenius triage "docs: typo" --repo org/repo --diff-stat "docs/faq.md | 3 ++-"

# Status heartbeat (outbound PR monitoring)
python3 -m prgenius status --author zsxh1990
python3 -m prgenius status --author zsxh1990 --format json --save-snapshot

# Profile writeback suggestions (dry-run)
python3 -m prgenius profile writeback --author zsxh1990

🤖 GitHub Action

Use PR Genius as a GitHub Action in any repo:

# .github/workflows/pr-genius.yml
name: PR Genius Check
on:
  pull_request:
    types: [opened, synchronize]

permissions:
  contents: read
  issues: write   # required when comment_mode != never (post/update PR comment)

jobs:
  pr-genius:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: zsxh1990/pr-genius@v1
        id: pr-genius
        with:
          title: ${{ github.event.pull_request.title }}
          repo: ${{ github.repository }}
          body: ${{ github.event.pull_request.body }}
          pr_number: ${{ github.event.pull_request.number }}
          comment_mode: always   # never | high_risk | always — post full analysis as a PR comment

comment_mode controls whether the full analysis is posted as a visible PR comment (mirrors pr-agent's /review):

• never — do not post (default when unset and comment_on_high_risk is false)
• high_risk — post only when the risk tier is high_risk
• always — post on every run; existing comments are updated in place (no spam)

Legacy: comment_on_high_risk: true is still supported and behaves like comment_mode: high_risk.

Version Auto-Update

• @v1 — Always points to the latest v1.x.x release (recommended)
• @v1.9.1 — Pinned to specific version (for reproducibility)
• @main — Latest development version (not recommended for production)

The v1 tag is automatically updated when a new version is published to PyPI.

Docker Image (Alternative)

Use PR Genius as a Docker container via GitHub Container Registry:

# .github/workflows/pr-genius-docker.yml
name: PR Genius Check (Docker)
on:
  pull_request:
    types: [opened, synchronize]

permissions:
  contents: read
  pull-requests: write

jobs:
  pr-genius:
    runs-on: ubuntu-latest
    steps:
      - name: Run PR Genius
        uses: docker://ghcr.io/zsxh1990/pr-genius:latest
        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
        with:
          args: coach "${{ github.event.pull_request.title }}" --repo ${{ github.repository }} --format json

Docker Image Tags:

• ghcr.io/zsxh1990/pr-genius:latest — Latest release
• ghcr.io/zsxh1990/pr-genius:1.9.1 — Specific version
• ghcr.io/zsxh1990/pr-genius:1.9 — Minor version
• ghcr.io/zsxh1990/pr-genius:1 — Major version (auto-updated)

Auto-update with Dependabot:

# .github/dependabot.yml
version: 2
updates:
  - package-ecosystem: "github-actions"
    directory: "/"
    schedule:
      interval: "weekly"
  - package-ecosystem: "docker"
    directory: "/"
    schedule:
      interval: "weekly"

🤖 MCP Configuration

{
  "mcpServers": {
    "pr-genius": {
      "command": "python",
      "args": ["-m", "prgenius", "mcp", "serve"]
    }
  }
}

Docker: docker run --rm -i ghcr.io/zsxh1990/pr-genius:1.9.1

12 MCP Tools

Tool Parameter Notes

• title (required for analyze_pr, coach_pr, triage_pr): The PR title, e.g. "fix: timeout in connection pool"
• repo (required for most tools): Repository in owner/name format, e.g. "encode/httpx"
• pr_description (optional): Additional PR body text for deeper analysis
• query (required for search_patterns): Search keywords, e.g. "connection timeout"

📊 Data Scale

按仓库规模分布

🤖 Robots / Agents

1. docs/index.md — file map
2. AGENT_GUIDELINES_SCHEMA.md — agent_guidelines schema
3. ROUNDS_SCHEMA.md — rounds schema
4. BLACKLIST.md — repos we don't track

📖 Contributing

See CONTRIBUTING.md. AI-assisted PRs welcome.

🤝 Community

• 📋 Code of Conduct
• 🔒 Security Policy
• 🐛 Issue Tracker
• 📜 Changelog

中文文档

中文版 README：README.zh-CN.md

Citation

@misc{pr-genius-2026,
  title  = {PR Genius — Evidence-backed PR Contribution Advisor},
  author = {zsxh1990},
  year   = {2026},
  url    = {https://github.com/zsxh1990/pr-genius}
}