debug
Investigation-first debugging — gather evidence, form confirmed root-cause hypothesis, hand off to fix mode with diagnosis file.…
Produce a frozen implementation plan with owned paths, configured gates, risks, and acceptance criteria.
$ npx -y skills add Borda/AI-Rig --skill rig-plan --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/rig-planContext preview
The summary Claude sees to decide when to auto-load this skill.
Produce a frozen implementation plan with owned paths, configured gates, risks, and acceptance criteria.
name: Rig Plan description: Produce a frozen implementation plan with owned paths, configured gates, risks, and acceptance criteria.
> Read [Evidence rules](../../references/evidence.md) before writing any factual claim.
Produce one plan that implementation can execute without re-deciding scope.
1. **Classify scope.** `TRIVIAL` under 3 files and 50 changed lines; `LOCAL` one subsystem; `BROAD` cross-subsystem or 8+ files; `HIGH_RISK` when auth, credentials, data, release, CI, or migrations are touched. State the classification and why. 2. **Collect evidence.** Read the actual source, tests, configuration, and consumers. Do not plan from filenames alone. Record each authoritative file in the run. 3. **Start the run.** Call `rig_evidence` with `action: "start"`, `skill: "rig-plan"`, and the target paths. Keep the returned `run_id`. 4. **Map the change.** List each affected component, its consumer, the owned path, and the check that proves it still works. 5. **Write the plan** to `<run-dir>/plan.md` with: goal, scope classification, numbered steps, owned paths, gates, risks, rollback, acceptance criteria, and explicit non-goals. 6. **Challenge the plan once.** Dispatch `rig-challenger` (native subagent when installed, otherwise `rig_dispatch`) with the plan, source digests, and acceptance criteria. Accept only findings that cite source. 7. **Freeze it.** Record the plan with `rig_evidence` (`action: "record"`, `plan_path`, and `reviewer` naming the independent challenger) and finalize.
Report the plan path, scope classification, digest, step count, gates, top risks, and any unresolved decision. Do not claim the change is implemented.
Practical agent workflows for Python, ML, and open-source maintenance. AI-Rig turns recurring work—scoping a change, reproducing a bug, reviewing a pull request, running an experiment, or checking release readiness—into explicit workflows with specialist
Investigation-first debugging — gather evidence, form confirmed root-cause hypothesis, hand off to fix mode with diagnosis file.…
TDD-first feature development — crystallise API as a demo test, drive implementation to pass it, run quality stack and…
Reproduce-first bug resolution — capture bug in failing regression test, apply minimal fix, run quality stack and review loop.…
Analysis-only planning — classify and scope a task without writing code; outputs a structured plan to .plans/active/. TRIGGER…
Test-first refactoring — audit coverage, add characterization tests, apply changes with safety net, run quality stack and review…
Multi-agent code review of local Python files, directories, or the current git diff covering architecture, tests, performance,…