why - Investigate why code and design decisions exist
Investigates design rationale, tradeoffs, constraints, regressions, and historical intent using parallel evidence searches with citations and calibrated confidence.
Tags
Updated: 2026-10-15Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Anchor questions in code
- Map available MCP connectors
- Search evidence categories in parallel
- Compare competing explanations
- Calibrate confidence explicitly
- Cite supporting sources
- Report evidence gaps
Inputs
- User investigation question
- Target code and file paths
- Git repository history
- GitHub pull request data
- Available MCP connectors
- Skill reference files
Outputs
- Cited rationale report
- Evidence coverage map
- Confidence assessment
- Competing explanations
- Evidence gaps and contradictions
Requirements
- Git repository access
- GitHub CLI access
- Codex-native subagent support
- Available MCP access for external evidence
- Access to skill reference files