agentsclimarketplace

Systematic debug

Skill IrfanSadiqRahat/constellation/methodology/systematic-debug

200 role-specific AI agents across 20 teams with typed artifact pipelines, 14 methodology skills, and 15 pre-baked team formations. The virtual engineering org for Claude Code, Cursor, Codex CLI.

Install
npx -y skills add IrfanSadiqRahat/constellation --skill systematic-debug

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Five-phase root cause process with role-aware diagnostics. Bisect, minimize, hypothesize, verify, document.

SKILL.md

2.7 KB, as published. Nobody here has run it

systematic-debug

Five phases (in order, no skipping)

1. Reproduce — observable

A bug not reproducible is a hypothesis, not a bug.

  • Minimal repro (delta-debugging if needed)
  • Environment captured (versions, env vars, data state)
  • Steps recorded so a peer can re-run

2. Isolate — narrow blast radius

  • Git bisect for regressions
  • Binary search on input space (works for data bugs)
  • Disable subsystems until repro disappears

3. Hypothesize — single hypothesis at a time

  • State the hypothesis as a falsifiable claim
  • Predict an observation that would confirm or refute
  • Run one experiment

4. Verify root cause — not symptom

A root cause is one whose removal makes the bug impossible, not just absent. If the same input could produce the same bug via a different path, you found a symptom.

5. Document — RootCauseReport

title: <bug>
severity: 1 | 2 | 3
repro: <minimal steps>
investigation_log: [<hypothesis, experiment, result, ts>]
root_cause: <statement>
mechanism: <how the cause leads to the bug>
fix: <change>
prevention:
  - test added: <link>
  - lint rule: <link>
  - design change: <link>
adjacent_risks: <where else this class of bug could live>

Role-aware diagnostics

rolediagnostic toolkit
react-specialistReact DevTools, <Profiler>, source maps, useDebugValue
frontend-perf-engineerRUM, traces, performance.mark, INP attribution
backend-architectdistributed tracing, exemplars, db EXPLAIN, profilers
postgres-specialistpg_stat_statements, EXPLAIN (ANALYZE, BUFFERS), auto_explain
data-architectlineage trace, sample-then-diff, freshness check
ai-engineertrace per request, eval bucket diff, prompt-version pin
incident-commandertimeline, comms log, dashboards as evidence
security-architecttamper-evident logs, IOC search, traffic capture

Forbidden moves

  • "It probably works now, retry" — confirmation bias, often hides flakiness
  • Multiple changes per experiment — can't attribute cause
  • "Works on my machine" closure — not closure
  • Fixing the symptom and moving on without root-cause analysis

Why this skill is opinionated

  • Five phases — explicit "isolate via bisect" phase
  • Role-aware toolkit per role — Postgres bugs ≠ React bugs ≠ AI hallucinations
  • Mechanism field, not just root cause
  • Adjacent risks surfaced
  • Prevention is tracked as required output

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.