IssuePilot automatically reviews incoming GitHub issues, identifies duplicates, extracts missing diagnostic context, and assigns priority labels—saving maintainers hours of manual triaging.
Built with a parallel-execution multi-agent architecture to analyze issues from multiple angles simultaneously.
Uses semantic similarity and context engineering to flag new issues that overlap with existing repository bugs or feature requests.
Automatically identifies if crucial information like environment details, reproduction steps, or logs are missing from the submission.
Assigns severity tiers (CRITICAL, HIGH, MEDIUM, LOW) based on impact and recommends maintainer response drafts and labels.
Google Agent Development Kit (ADK) orchestrates specialized AI agents.
GitHub Webhooks push new issues into the PostgreSQL database.
Parallel ADK agents run evaluations using Gemini 2.5 Flash.
Labels and priorities are synced back to the UI dashboard.