incident-to-regression

What problem does it solve?

Incident artifacts often exist as ad-hoc evidence that is hard to reproduce in CI. This skill converts those artifacts into deterministic regression fixtures and CI-ready outputs that can be replayed and validated.

Core Features & Use Cases

• Initialize deterministic fixtures from an incident source using gait regress init.
• Parse initialization outputs to extract fixture metadata and next steps.
• Execute regression graders to validate consistency and produce CI-friendly reports (e.g., JUnit).
• Use cases include incident triage, reproducible failure reproduction, and automated evidence generation for PRs and audits.

Quick Start

Provide a run_source (run id, runpack path, or equivalent) and a writable workdir, then run gait regress init to create deterministic fixtures, followed by gait regress run to generate regression outputs for CI.