How does the priority scoring algorithm work?

It uses a weighted formula that prioritizes P-level labels (P0-P3) and adjusts scores based on keywords like 'production', 'security', or 'regression' found in the issue description.

What happens if the AI subagent is unable to fix a specific bug?

The skill will clean up the temporary branch, post a detailed comment on the issue with its findings and blockers, and apply a 'needs-review' label for manual intervention.

Can I limit the number of issues processed in a single run?

Yes, you can use the --max-issues parameter or specific priority flags like --p0-only to control exactly how many issues the sweep should attempt to resolve.

Does it run tests before submitting a fix?

Yes, it requires a Docker environment and executes specific test commands to ensure the fix works and doesn't introduce regressions before the PR is created.

Can I use this skill to fix bugs directly on the main branch?

No, the skill is strictly configured to branch from and target the 'development' branch to maintain standard integration workflows and protect production-ready code.

GitHub Bug Sweep

Name: GitHub Bug Sweep
Author: arlenagreer

byarlenagreer

•

开发者工具

Orchestrates the automated identification, prioritization, and resolution of GitHub issues using sequential AI subagents.

GitHub Bug Sweep is a high-level orchestration workflow designed for systematic bug resolution within repositories following the development branch model. It automatically prioritizes open issues from P0 to P3 based on a weighted scoring algorithm, deploys dedicated subagents to investigate and fix each bug, and manages the entire lifecycle from branch creation to pull request generation. This skill is perfect for clearing technical debt, handling maintenance sprints, and ensuring that critical security vulnerabilities are addressed immediately with comprehensive documentation and verified test results.

主要功能

01Intelligent priority scoring (P0-P3) based on labels and issue body analysis

02Sequential subagent deployment for isolated bug investigation and repair

03Dry-run mode for previewing prioritization and estimation without side effects

04Strict branch workflow management targeting the development branch exclusively

052 GitHub stars

06Automated PR creation including root cause analysis and test verification

使用场景

01Automating the repetitive manual steps of branch setup, bug fixing, and PR documentation

02Systematically clearing a backlog of accumulated bug reports during maintenance periods

03Rapidly addressing high-priority security vulnerabilities and regressions

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add arlenagreer/claude_configuration_docs github-bug-sweep

For use in Claude.ai and ChatGPT

Download Skill