agentleFS
vibe-cokitvibe-cokit/claude-code/skills/vk-debug/SKILL.md

vk-debug

Skill1 starsChanged 4 months agoApache-2.0View on GitHub

⚡⚡ Trace the root cause of a reported issue and return an evidence-backed diagnosis (no auto-fix).

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If a critical signal is missing (no error text AND no reproduction), ask ONE concise clarifying question before investigating (per Decision …

    line 19
  2. Very specific:

    If intake (§1) surfaces ≥2 independent hypotheses (e.g.

    line 33
  3. Distinctive:

    IMPORTANT: Do NOT implement the fix automatically.

    line 54
  4. Distinctive:

    IMPORTANT: No root-cause claim without fresh, concrete evidence.

    line 55

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Debug — Root Cause Investigation

Reported current state: $ARGUMENTS

Take the user's current project state (what's broken, errors, symptoms, when it started, recent changes) and return the ROOT CAUSE — the origin where the defect was introduced — backed by concrete evidence. Stop at the diagnosis; do NOT implement a fix.

1. Intake

Parse $ARGUMENTS into: symptom, error text/stack, affected component & scope, timeframe (when it started), recent changes (deploys/commits), reproduction steps.

  • If a critical signal is missing (no error text AND no reproduction), ask ONE concise clarifying question before investigating (per Decision Classification in orchestration-protocol.md). Otherwise proceed.

2. Investigate (orchestrate)

Delegate the investigation to the debugger subagent. Instruct it to follow the debugging skill Four-Phase process:

  1. Root Cause Investigation — read errors, reproduce, check recent changes, gather evidence.
  2. Pattern Analysis — compare against working examples.
  3. Hypothesis & Testing — form a theory, test minimally.
  4. Backward root-cause-tracing — trace the symptom backward level-by-level through the call stack to the origin where invalid data/state was introduced; the fix-point is the source, not the symptom. Pass the intake fields (and the plan reports dir, if running inside one) to the subagent.

If intake (§1) surfaces ≥2 independent hypotheses (e.g. two unrelated suspect components), spawn one debugger instance per hypothesis in parallel (≤4, read-only — no ownership conflict possible; see workflows/parallel-execution.md §4). Single clear hypothesis → one instance, unchanged.

3. Verify

The debugger's report is an UNVERIFIED claim (Verification Gate, development-rules.md). Confirm the root cause is backed by concrete evidence — log excerpts, stack/trace, code refs (file:line), or a reproduction — not speculation. Reject symptom-level conclusions.

4. Report

Return a concise root-cause report:

  • Root cause — 1-2 lines: the origin defect.
  • Evidence chain — symptom → trace-back steps → origin (with file:line, logs).
  • Confidence — high / medium / low. If not definitive, list ranked likely causes + next investigation steps.
  • Suggested remediation direction — where & how to fix conceptually (NO code). May suggest /vk-plan or /vk-fix as the next step. Save as an inter-agent report in plans/<plan-name>/reports/ only when running inside a plan dir; otherwise present inline.

IMPORTANT: Do NOT implement the fix automatically. IMPORTANT: No root-cause claim without fresh, concrete evidence. IMPORTANT: Activate the debugging skill; /vk-scout to locate files.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.