What problem does it solve?
Standard code search often retrieves irrelevant or unstable code because it ignores the historical context of the codebase. This skill solves the problem of agents copying bug-prone, legacy, or siloed code by reordering search results based on git history and structural signals.
Core Features & Use Cases
- Preset-based Reranking: Apply pre-configured logic for common tasks like identifying hotspots, tech debt, or security-sensitive legacy code.
- Custom Weight Composition: Build bespoke ranking models using specific signals like bug-fix rates, ownership, and churn to suit unique project needs.
- Use Case: When performing a security audit, use this skill to prioritize old code in authentication paths that has a high history of bug fixes, ensuring the agent focuses on the most critical areas.
Quick Start
Invoke the analytics-rerank skill to apply the hotspots preset to your current search results for identifying bug-prone areas.