AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Dev Debug

skill-christopherlouet-claude-base-dev-debug · by christopherlouet

Debug and resolve problems. Use when the user has a bug, an error, an unexpected behavior, or wants to understand why something is not working.

No reviews yet
0 installs
13 views
0.0% view→install

Install

$ agentstack add skill-christopherlouet-claude-base-dev-debug

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-christopherlouet-claude-base-dev-debug)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
1mo ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Dev Debug? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Debug a Problem

Goal

Identify the root cause of a bug methodically via a systematic 4-phase approach.

Systematic Methodology (4 Phases)

┌──────────────────────────────────────────────────────────────────┐
│                   SYSTEMATIC DEBUGGING                            │
├──────────────────────────────────────────────────────────────────┤
│                                                                   │
│  PHASE 1: OBSERVATION          Collect without interpreting       │
│  ═══════════════════                                              │
│  - Reproduce the exact symptom                                    │
│  - Document the environment                                       │
│  - Capture logs, stack traces, states                             │
│  - DO NOT jump to conclusions                                     │
│                                                                   │
│  PHASE 2: HYPOTHESES           Reason systematically              │
│  ════════════════════                                             │
│  - List ALL possible causes                                       │
│  - Rank by probability (high/medium/low)                          │
│  - Define a validation test for each hypothesis                   │
│  - Use the 5 Whys technique                                       │
│                                                                   │
│  PHASE 3: INVESTIGATION        Prove, do not assume               │
│  ══════════════════════                                           │
│  - Test ONE hypothesis at a time                                  │
│  - Use strategic tracing and logging                              │
│  - Isolate with binary search (code or git bisect)                │
│  - Document each tested hypothesis (confirmed/refuted)            │
│                                                                   │
│  PHASE 4: VERIFICATION         Confirm the fix is real            │
│  ════════════════════                                             │
│  - Reproduce the original bug (must fail without the fix)         │
│  - Apply the minimal fix                                          │
│  - Prove the bug is fixed                                         │
│  - Verify the absence of side effects                             │
│  - Add a non-regression test                                      │
│  - Defense in depth: assertions on invariants                     │
│                                                                   │
└──────────────────────────────────────────────────────────────────┘

Phase 1: Observation

Key questions:

  • What is happening exactly?
  • What should happen?
  • When did it start?
  • Is it 100% reproducible?
  • What are the aggravating/mitigating factors?
# Recent logs
tail -100 logs/app.log 2>/dev/null

# Latest commits
git log --oneline -10

# Recent changes in the suspect area
git log --oneline -5 -- src/path/suspect/

Phase 2: Hypotheses

Hypothesis matrix

| # | Hypothesis | Probability | Validation test | |---|------------|-------------|-----------------| | 1 | [Most likely] | High | [How to verify] | | 2 | [Secondary] | Medium | [How to verify] | | 3 | [Less likely] | Low | [How to verify] |

5 Whys technique

Problem: The application crashes at login

1. Why? -> The JWT token is invalid
2. Why? -> The token has expired
3. Why? -> The refresh token was not called
4. Why? -> The interceptor did not detect the expiration
5. Why? -> Timezone bug in the comparison

Root cause: Timezone bug in the refresh logic

Common causes by type

| Bug type | Frequent causes | |----------|-----------------| | Null/Undefined | Missing data, race condition, API changed | | Type error | Wrong type, JSON parsing, implicit conversion | | Off-by-one | Array index, loop, ` The bug is in the remaining half

  • No -> The bug is in the commented half
  1. Repeat until isolating the exact line

### Git bisect (find the faulty commit)

```bash
git bisect start
git bisect bad                 # Current version broken
git bisect good        # Last known good version
# Test and mark good/bad until finding the commit
git bisect reset

Phase 4: Verification (MANDATORY)

Prove the fix works

1. WITHOUT the fix: reproduce the bug -> failure confirmed
2. WITH the fix: same scenario        -> success confirmed
3. Existing tests                     -> all pass
4. Non-regression test                -> written and passes
5. Side effects                       -> verified absent

Defense in depth

// Add assertions on critical invariants
function processPayment(amount: number, userId: string) {
  assert(amount > 0, 'Payment amount must be positive');
  assert(userId, 'User ID is required');
  // ...business code...
}

Completion checklist

[ ] Bug reproduced reliably
[ ] Root cause identified (not just the symptom)
[ ] Minimal fix applied (no opportunistic refactoring)
[ ] Non-regression test added
[ ] Existing tests pass
[ ] No side effects
[ ] Fix documentation (descriptive commit message)

Expected output

## Diagnosis: [Bug description]

### Phase 1 - Observation
**Symptom:** [What happens]
**Expected behavior:** [What should happen]
**Reproduction:** [Steps 1, 2, 3...]

### Phase 2 - Hypotheses
| # | Hypothesis | Probability | Result |
|---|------------|-------------|--------|
| 1 | [...] | High | Confirmed/Refuted |

### Phase 3 - Investigation
**Root cause:** `src/xxx.ts:42` - [Technical explanation]
**5 Whys:** [Causal chain]

### Phase 4 - Verification
- [x] Bug reproduced
- [x] Fix applied
- [x] Non-regression test added
- [x] All tests pass
- [x] No side effects

Rules

  • Do not assume - verify (Phase 4 mandatory)
  • One bug at a time
  • Understand BEFORE fixing
  • Always add a non-regression test
  • Document every tested hypothesis, even refuted ones
  • The fix must be MINIMAL - no opportunistic refactoring

Iron Law: No fix without root cause

IMPORTANT: NEVER propose a fix before identifying the root cause. Symptom fixes mask the real problem and create new bugs.

Red flags (rationalizations to avoid)

| Phrase | Problem | |--------|---------| | "Quick fix for now" | Avoids the root cause | | "Let's just try changing X" | Guess-and-check, not systematic | | "I don't really understand but it should work" | Blind fix | | "It's urgent, no time to investigate" | Systematic investigation is FASTER |

Rule of 3 failures

After 3 failed fix attempts: STOP. Do not attempt a 4th fix.

  1. Question the basic assumptions
  2. Broaden the search perimeter
  3. Check whether the problem is architectural (not just a local bug)
  4. Consider git bisect to find the faulty commit

Metrics

| Approach | Average time | 1st-attempt fix rate | |----------|--------------|----------------------| | Systematic (4 phases) | 15-30 min | ~95% | | Random fixing (guess-and-check) | 2-3h | ~40% |

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.