Agent-Skills.md

Agent Skills: Debugger

Systematic debugging skill. 7-step workflow: Reproduce, Minimize, Hypothesize, Instrument, Fix, Prevent, Verify. Activate when troubleshooting errors.

UncategorizedID: ydnikolaev/antigravity-factory/debugger

Repository

ydnikolaev/antigravity-factory
ydnikolaev
83

Install this agent skill to your local

pnpm dlx add-skill https://github.com/ydnikolaev/antigravity-factory/tree/HEAD/src/skills/debugger

Skill Files

Browse the full folder contents for debugger.

Download Skill

Loading file tree…

src/skills/debugger/SKILL.md

Skill Metadata

Name
debugger
Description
"Systematic debugging skill. 7-step workflow: Reproduce, Minimize, Hypothesize, Instrument, Fix, Prevent, Verify. Activate when troubleshooting errors."

Debugger

[!IMPORTANT]

First Step: Read Project Config & MCP

Before making technical decisions, always check:

| File | Purpose | |------|---------| | project/CONFIG.yaml | Stack versions, modules, architecture | | mcp.yaml | Project MCP server config | | mcp/ | Project-specific MCP tools/resources |

Use project MCP server (named after project, e.g. mcp_<project-name>_*):

  • list_resources → see available project data
  • *_tools → project-specific actions (db, cache, jobs, etc.)

Use mcp_context7 for library docs:

  • Check mcp.yaml → context7.default_libraries for pre-configured libs
  • Example: libraryId: /nuxt/nuxt, query: "Nuxt 4 composables"

When to Activate

  • Runtime errors, crashes
  • Failing tests
  • "It used to work" regressions
  • Unexpected behavior
  • Performance/timeout issues (initial triage)

The 7-Step Debug Workflow

[!CAUTION] DO NOT SKIP STEPS! Each is critical for systematic debugging.

Step 1: Reproduce (Test Cases First) 🔁

MANDATORY: You must create a reproduction artifact.

  • Unit Test: If logic error, write a failing unit test.
  • Script: If integration error, write a standalone repro script.
  • Config: If environment error, document exact config to repro.

Status Check: Do you have a "Red" test? If no, go back.

Step 2: Minimize 🎯

Reduce to smallest repro:

  • One file
  • One function
  • Smallest dataset
  • Remove unrelated code

Step 3: Hypothesize 🧠

Form 2-5 hypotheses, ranked by likelihood:

  1. Most likely cause
  2. Second candidate
  3. Edge case possibility
  4. (Optional) Unlikely but possible

Step 4: Instrument 🔧

Add temporary debugging:

  • Logging statements
  • Assertions
  • Breakpoints
  • Print variables at key points

Use existing diagnostics if available.

Step 5: Fix ⚡

Apply smallest change that removes root cause:

  • Don't over-engineer
  • Fix the actual problem, not symptoms
  • Keep changes minimal

Step 6: Prevent 🛡️

Add protection:

  • Regression test
  • Validation guard
  • Assertion
  • Error handling improvement

Step 7: Verify ✅

Run verification:

  • The failing case now passes
  • Related test suites pass
  • No new regressions

Report Format

When reporting a fix, use this structure:

### Symptom
(What was wrong)

### Repro Steps
1. ...
2. ...

### Root Cause
(Why it happened)

### Fix
(What you changed)

### Regression Protection
(Test or guard added)

### Verification
- Commands run:
- Results:
<!-- INCLUDE: _meta/_skills/sections/language-requirements.md -->

Team Collaboration

  • Backend: @backend-go-expert (You debug their code)
  • Frontend: @frontend-nuxt (You debug their code)
  • QA: @qa-lead (They report issues to you)

When to Delegate

  • ✅ Delegate to @qa-lead when: Fix is complete, needs testing
  • ⬅️ Return to reporter when: More info needed to reproduce
  • 🤝 Coordinate with code owner when: Fix requires architectural changes
<!-- INCLUDE: _meta/_skills/sections/brain-to-docs.md -->

Document Lifecycle

Protocol: DOCUMENT_STRUCTURE_PROTOCOL.md

| Operation | Document | Location | Trigger | |-----------|----------|----------|---------| | 🔵 Creates | <issue-name>.md | active/bugs/ | Debug report complete | | 📖 Reads | Issue description, logs | — | On activation | | 📖 Reads | Implementation docs | active/backend/, active/frontend/ | Understanding context | | 📝 Updates | ARTIFACT_REGISTRY.md | project/docs/ | On create, on complete | | 🟡 To Review | <issue-name>.md | review/bugs/ | Fix verified | | ✅ Archive | — | closed/bugs/<id>/ | @doc-janitor on final approval |

Pre-Handoff Validation (Hard Stop)

[!CAUTION] MANDATORY self-check before notify_user or delegation.

| # | Check | |---|-------| | 1 | ## Upstream Documents section exists with paths | | 2 | ## Requirements Checklist table exists | | 3 | All ❌ have explicit Reason: ... | | 4 | Document in review/ folder | | 5 | ARTIFACT_REGISTRY.md updated |

If ANY unchecked → DO NOT PROCEED.

Handoff Protocol

[!CAUTION] BEFORE handoff:

  1. Save final document to project/docs/ path
  2. Change file status from Draft to Approved in header/frontmatter
  3. Update project/docs/ARTIFACT_REGISTRY.md status to ✅ Done
  4. Use notify_user for final approval
  5. THEN delegate to next skill

Tech Debt Protocol (Hard Stop)

[!CAUTION] Follow ../standards/TECH_DEBT_PROTOCOL.md. When creating workarounds:

  1. Add // TODO(TD-XXX): description in code
  2. Register in project/docs/TECH_DEBT.md

Forbidden: Untracked TODOs, undocumented hardcoded values.

Git Protocol (Hard Stop)

[!CAUTION] Follow ../standards/GIT_PROTOCOL.md.

  1. Branch: Create fix/<bug-name> branch before fixing.
  2. Commit: Use fix(<scope>): <description> format.
  3. Atomic: One fix = One commit (regression test included).

Reject: "wip", "debug", "fixed" as commit messages.

Antigravity Best Practices

  • Use task_boundary for multi-step debugging sessions
  • Use notify_user to confirm root cause before fixing
  • Always add regression tests