bug
Bug Investigation and Fix
Structured bug workflow: capture → reproduce → root cause → failing test → fix → verify.
Iron Laws
- Reproduce before investigating. You cannot fix a bug you cannot trigger reliably.
- Find root cause before writing any fix. Symptom fixes are failure. Three failed fix attempts means you haven't found the real cause yet.
- Write a failing test before the fix. The test proves your diagnosis is correct and prevents regression.
- One hypothesis at a time. Test the most likely cause with the smallest possible change. Never change multiple things simultaneously.
- Verify with the original repro — not just the unit test. The user-visible symptom must be confirmed resolved.
Phase 1 — Capture
Collect everything needed to understand and reproduce the issue:
Required information:
More from supatest-ai/supa-skills
debug-prod
Investigate a production incident or live bug. Follows mitigation-first protocol — stops the bleeding before investigating, then uses the observability funnel (metrics → traces → logs) to find root cause. Produces a structured findings report and optional postmortem.
1agent-readiness
Evaluate how ready a codebase is for autonomous AI agent work. Produces a scored report across 11 dimensions with a maturity level and prioritized action plan.
1work-summary
Generate a summary of work done based on git commit history across one or more repositories. Use for standups, timesheets, retrospectives, or any time you need a structured overview of what was accomplished in a time period.
1test-feature
Interactively test a completed feature using browser automation. Navigates the live app, exercises all key flows, captures a GIF video of the happy path, and produces a visual verification report. Use after implementing a new feature or making code changes that need testing verification.
1review-pr
Review a pull request or set of code changes. Filters to issues introduced by this PR (not pre-existing), scores confidence, and produces a prioritized report with severity tiers. Requires user approval before posting anything to GitHub.
1signoz-health-check
Check broad metrics and health status of SigNoz observability platform. Shows service health, error rates, latency, recent errors, and alerts. Use when you need to monitor system health or investigate issues.
1