An AI-powered pull request assistant that identifies duplicate submissions, scores contribution quality, and generates prioritized triage reports.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install pr-triage
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install pr-triage using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is PR Triage?
The PR Triage skill is a specialized automation tool designed for maintainers of high-traffic repositories. By integrating with Openclaw Skills, this agent streamlines the review process by analyzing open Pull Requests (PRs) to detect redundant work and assess the overall health of contributions. It eliminates the manual effort required to sort through hundreds of submissions by providing objective data on which PRs are most ready for human review.
This skill uses sophisticated similarity detection and quality scoring algorithms to help teams maintain repository standards. It ensures that maintainers can focus on high-impact code changes while providing clear feedback to contributors regarding duplicate or low-quality submissions. Using Openclaw Skills for triage significantly reduces the time-to-merge for critical bug fixes and features.
PR Triage Use Cases
- Managing large-scale open-source projects with high PR volume.
- Automatically flagging duplicate PRs that target the same issue or files.
- Prioritizing code reviews based on automated quality scoring (tests, descriptions, size).
- Auditing stale pull requests to clean up repository backlogs.
- Generating weekly maintainer reports to visualize contributor activity and PR health.
How PR Triage Works
- Metadata Acquisition: The skill utilizes the GitHub CLI to fetch open PR metadata, including titles, bodies, file paths, and labels.
- Intent Extraction: It normalizes PR data by extracting key intent signals such as function names, error messages, and issue references.
- Similarity Analysis: The agent calculates a similarity score between PRs using Jaccard similarity of changed files and keyword overlap.
- Quality Grading: Each PR is assigned a score based on objective signals like the presence of tests, description length, and commit size.
- Report Generation: A detailed Markdown report is generated, grouping duplicates and ranking PRs by their quality grade (A-D).
- Automated Interaction: Optionally, the agent can post comments or add labels to GitHub to facilitate communication with contributors.
PR Triage Setup
To use this skill, ensure you have the GitHub CLI installed and authenticated. Within your Openclaw Skills environment, you can run the triage command with various flags:
# Analyze PRs from the last 7 days in a specific repo
/pr-triage --repo owner/repo --days 7
# Run a full audit and save the results to a file
/pr-triage --repo owner/repo --all --output triage-report.md
PR Triage Data Schema & Taxonomy
The skill processes and generates data based on the following taxonomy:
| Attribute | Description |
|---|---|
| Similarity Score | A weighted metric (0-100) combining file overlap (60%) and keyword similarity (40%). |
| Quality Signals | Points awarded for tests, descriptions, issue references, and PR size. |
| Intent Mapping | Extracted keywords, file paths, and action verbs from PR content. |
| Duplicate Groups | Collections of PRs that reference the same issue or exceed the similarity threshold. |
PR Triage Advanced Features
- Multi-signal duplicate detection including file overlap and natural language intent.
- Automated labeling support for 'duplicate' and 'needs-review' tags via Openclaw Skills.
- Customizable similarity thresholds to adjust the sensitivity of duplicate flagging.
- Intelligent contributor context checking to identify first-time vs. recurring authors.
- Token-optimized workflows that prioritize metadata analysis over expensive diff reading.
SKILL.md
Loading
Related Openclaw Skills
Interesting Finding Link Digest
An automated workflow that transforms shared Discord links into analyzed summaries and structured local knowledge base entries.
Interesting Finding
An automated workflow that transforms shared Discord links into structured, categorized knowledge base entries with AI-generated summaries.
GitHub Cred
A comprehensive scoring engine that quantifies GitHub developer influence and contribution quality through a four-tier metric system.
gh-standup
Generate concise GitHub standup reports by automatically aggregating merged pull requests and direct commits from repositories or organizations.
Raindrop Bookmark Sync
Sync and automate the ingestion of Raindrop.io bookmarks into your AI-driven knowledge base using Openclaw Skills.
hume-x
A comprehensive X/Twitter CLI and SDK leveraging GraphQL and cookie-based authentication for full platform engagement and automation.
Featured*
