Test Failure Investigator

Runbook-style skill for systematic test failure investigation. Given a failing test, determines root cause and recommends action.

Activation

/test-failure-investigator [test-name-or-file]

Investigation Flow

Step 1: Classify the Failure

Run the test 3 times and classify:

# Run test 3 times
for i in 1 2 3; do npx jest {{test_file}} 2>&1 | tail -5; echo "--- Run $i ---"; done

Step 2: Narrow the Scope

# When did it start failing?
git log --oneline -20 -- {{related_source_files}}

# What changed recently?
git diff HEAD~5 -- {{related_source_files}}

# Does it fail in isolation?
npx jest {{test_file}} --testNamePattern="{{test_name}}"

# Does it fail with other tests?
npx jest --runInBand  # sequential execution

Step 3: Root Cause Analysis

Step 4: Flaky Test Investigation

# Run 10 times to confirm flakiness
for i in $(seq 1 10); do npx jest {{test_file}} --forceExit 2>&1 | grep -E 'PASS|FAIL'; done

# Common flaky causes:
# - Shared state between tests (missing cleanup)
# - Time-dependent assertions (use fake timers)
# - Race conditions (missing await)
# - Port conflicts (use random ports)
# - Order dependency (run with --randomize)

Step 5: Report

## Test Failure Report
- **Test**: {{test_name}}
- **File**: {{test_file}}
- **Classification**: Regression / Flaky / Environment / Transient
- **Root Cause**: {{description}}
- **First Failed**: {{commit_hash}} ({{date}})
- **Fix**: {{recommended_action}}
- **Verified**: [ ] Fix applied and test passes 3x consecutively

Composition

After investigation, compose with:

• /bug-reporting-excellence — if regression found, file a bug report
• /regression-testing — if regression, add to regression suite
• /qe-test-execution — for re-running tests after fix

Gotchas

• Agent may guess at root cause without running the test — always reproduce first
• "Works on my machine" is not a diagnosis — compare environments (node version, OS, deps)
• Flaky tests that pass 9/10 times will still be reported as "passing" by CI — run 10+ times
• Test isolation failures are the #1 cause of flaky tests — check for shared state in beforeAll/afterAll