What problem does it solve?
Manually managing GitHub issue workflows, tracking triage status, and preparing well-specified tasks for AI agents is disjointed and inefficient for software development teams, leading to missed issues, inconsistent labeling, and wasted time on underspecified requests.
Core Features & Use Cases
- Label-based state machine: Enforces consistent issue workflow rules with clear transitions between triage states like needs-triage, ready-for-agent, and wontfix.
- Interactive grilling sessions: Guides maintainers through targeted questions to gather missing context from issue reporters, ensuring no critical details are overlooked.
- Durable agent brief generation: Creates standardized, long-lasting specifications for AFK agents that remain useful even as the codebase changes over time.
- Out-of-scope knowledge base: Tracks rejected feature requests to avoid duplicate discussions and preserve institutional decision-making context.
- Use case: A maintainer with 100+ open GitHub issues can use this skill to quickly prioritize work, hand off well-defined bugs to AI agents for implementation, and close duplicate feature requests with documented reasoning.
Quick Start
Use the github-triage skill to triage all unlabeled issues in your current GitHub repository and generate a prioritized list of items needing maintainer attention.