bugmagnet

BugMagnet

Systematic test coverage analysis and bug discovery.

Based on gojko/bugmagnet-ai-assistant.

Workflow

🔍 ANALYSE  → Understand code and existing tests
📊 GAP      → Identify missing coverage
✍️ WRITE    → Implement tests iteratively
🔬 ADVANCED → Deep edge case exploration (optional)
📋 SUMMARY  → Document findings and bugs

Interactive mode (default): STOP and wait for user confirmation between phases. Autonomous mode: When invoked by another agent (e.g. code-reviewer), skip all STOP points and proceed through all phases automatically. Present the full summary at the end instead of pausing for input.

---

Phase 1: Analysis (🔍 ANALYSE)

1. Detect language and testing conventions
2. Read implementation — public API, parameters, state, dependencies
3. Locate test file — if none exists, ask user about creating one
4. Run baseline coverage if tools available
5. Read existing tests — understand current coverage and patterns
6. Ask user: "Are there additional files I should review?"

STOP — Wait for user input. (Skip in autonomous mode.)

---

Phase 2: Gap Analysis (📊 GAP)

Evaluate missing coverage using edge-cases.md:

• Boundary conditions, error paths, state transitions
• Complex interactions, domain-specific edge cases
• Violated domain constraints

Categorise: High (core, errors, boundaries) → Medium (interactions, state) → Low (rare, performance). Present analysis with specific examples.

STOP — Ask user which tests to implement. (Skip in autonomous mode — implement all high-priority tests.)

---

Phase 3: Test Implementation (✍️ WRITE)

For each test, highest priority first:

1. Write single test (or 2-3 related)
2. Name describes outcome: "returns X when Y", "throws error when Z"
3. Run immediately

On failure: test expectation wrong → update test. Bug discovered → create skipped test with documentation. Need more context → try 2 more variations.

When bug found: create minimal reproduction, explore surrounding territory (bugs cluster), document in skipped test — DO NOT FIX, only document. Include: brief description, root cause, code location, expected vs actual, proposed fix.

Maximum 3 attempts per test — document and move on.

STOP — Ask user if they want advanced coverage. (Skip in autonomous mode — proceed to advanced if high-priority gaps remain.)

---

Phase 4: Advanced Coverage (🔬 ADVANCED)

Create separate test suite: "bugmagnet session <date>"

Work through edge-cases.md comprehensively — complex interactions, error handling, numeric/date/string/collection edge cases, state transitions, domain-specific cases.

---

Phase 5: Summary (📋 SUMMARY)

## Test Coverage Summary

**Tests Added:** X total (Y passing, Z skipped/bugs)

**Bugs Discovered:**
1. Bug name — file:line — root cause — proposed fix

Reference Files

• Edge Case Checklist