Help us improve
Share bugs, ideas, or general feedback.
From c3
Systematic bug fixing with TDD approach. Use when fixing bugs, debugging issues, or investigating problems. Accepts bug descriptions in any format, coordinates analyst/reviewer agents, creates tests before fixes, produces analysis reports.
npx claudepluginhub christophevg/marketplace --plugin c3How this skill is triggered — by the user, by Claude, or both
Slash command
/c3:bug-fixingThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
A systematic, test-driven workflow for fixing software bugs with agent coordination.
Guides technical evaluation of code review feedback: read fully, restate for understanding, verify against codebase, respond with reasoning or pushback before implementing.
Share bugs, ideas, or general feedback.
A systematic, test-driven workflow for fixing software bugs with agent coordination.
| Capability | Description |
|---|---|
| Bug Intake | Accept text, issue references, or structured bug reports |
| Analysis | Root cause investigation with functional analyst review |
| TDD Approach | Failing test first, then fix implementation |
| Agent Coordination | Orchestrate functional-analyst, ui-ux-designer, code-reviewer |
| Documentation | Bug analysis reports with issue comments |
Use this skill when:
Parse bug description → Detect project context → Extract details → Assign ID
Input Formats:
Project Detection:
| Detect | Method |
|---|---|
| Language | pyproject.toml, package.json, Cargo.toml |
| Framework | Config files (Django, React, Vue) |
| Test framework | pytest.ini, jest.config.js, vitest.config.ts |
| Conventions | .prettierrc, pylintrc, ruff.toml |
Invoke functional-analyst agent to:
Create bug analysis report:
docs/bug-analysis/{bug-id}.mdAnalyst outcomes:
| Outcome | Action |
|---|---|
| Confirmed (no UI) | Proceed to Phase 3 |
| Confirmed (with UI) | Proceed, note UI review needed |
| Rejected | Document reason, close bug |
Apply systematic debugging framework:
| Step | Action |
|---|---|
| Isolate | Reproduce consistently, identify boundaries |
| Gather Info | Log strategically, compare working vs broken |
| Hypothesize | Specific testable hypotheses, one variable at a time |
| Validate | Run tests, document findings |
RCA Techniques:
Critical: Create failing test BEFORE implementing fix.
Determine test type:
| Bug Type | Test Type |
|---|---|
| Logic/validation | Unit test |
| Integration/API | Integration test |
| User flow | E2E test |
Create test that demonstrates bug:
# Example: Test expects current (incorrect) behavior
def test_login_button_disabled():
result = login_button.is_enabled()
assert result == False # Passes, proving bug exists
Run test to confirm reproduction
Functional Analyst Review
↓ Approved
UI Changes? → UI/UX Designer Review
↓ Approved
Code Reviewer Review
↓ Approved
Proceed to Phase 7
Rejection Handling:
| Scenario | Max Iterations | Escalation |
|---|---|---|
| Analyst rejects | 2 rounds | Ask user |
| Tests fail | 3 attempts | Ask user |
| Cannot reproduce | 1 request | Close bug |
⚠️ Bug fix is NOT complete until CI passes.
After creating the PR, MUST:
Assign and Request Review:
gh pr edit {number} --add-assignee {user}
gh pr edit {number} --add-reviewer {user}
Check CI status:
gh pr checks {number}
Wait for CI to complete (poll if needed)
If CI fails:
gh run view {id} --log-failedOnly report bug fix complete when CI passes
See patterns/bug-analysis-template.md for the full template.
Key sections:
See patterns/test-creation-patterns.md for platform-specific guidance.
Common bug categories:
| Platform | Common Causes | Debug Focus |
|---|---|---|
| Frontend | Async race conditions, stale state | Timeline, state changes |
| Backend | N+1 queries, connection exhaustion | Query patterns, logs |
| Mobile | Device/OS variations, memory leaks | Environment, profiling |
| Database | Missing indexes, stale statistics | Query plans, metrics |
This skill does NOT handle:
| Issue | Solution |
|---|---|
| Cannot reproduce | Request more info, check environment differences |
| Tests keep failing | Analyze if fix incomplete or test incorrect |
| Multiple fix proposals | Analyst recommends, user decides |
| Agent rejects fix | Iterate with feedback (max 2 rounds) |
patterns/bug-analysis-template.md - Bug analysis report templatepatterns/test-creation-patterns.md - Platform-specific test patternspatterns/rca-techniques.md - Root cause analysis techniques