AgentSkillsCN

Analyze Test Results

从 GitHub Actions 下载测试结果、截图与日志

SKILL.md
--- frontmatter
name: Analyze Test Results
description: Analyze test failures and CI artifacts to identify and fix bugs
allowed-tools:
  - Bash
  - Read
  - Grep
  - Glob
  - Edit
context: auto

Analyze Test Results Skill

Analyze test results from CI or local runs to identify failures, diagnose root causes, and fix bugs.

When to Use

  • •After tests fail (CI or local)
  • •After running /download-ci-artifacts
  • •When asked to "fix test failures"
  • •When asked to "analyze test results"

Workflow

1. Locate Test Results

bash
# From CI artifacts
cat ci-artifacts/test-report/test-report.md

# From local run
uv run pytest MouseMaster/Testing/Python/ -v --tb=long 2>&1 | tee test-output.log

2. Parse JUnit XML (if available)

bash
# Extract failed tests from XML
grep -E "(failure|error)" ci-artifacts/unit/unit-tests.xml | head -20

3. For Each Failure

a. Identify the failing test

code
FAILED MouseMaster/Testing/Python/test_event_handler.py::TestClass::test_name

b. Read the test file

Understand what the test expects:

  • •What behavior is being tested?
  • •What are the assertions?
  • •What mocks are set up?

c. Read the implementation

Find the code being tested and understand:

  • •Current behavior
  • •Why it might fail
  • •Edge cases

d. Diagnose root cause

Common issues:

  • •Mock not configured: Check mock setup in test
  • •API changed: Update test or implementation
  • •Race condition: Add proper synchronization
  • •Missing dependency: Check imports and fixtures

4. Fix the Issue

Apply the minimal fix needed:

python
# If test expectation is wrong
# → Update the test

# If implementation is wrong
# → Fix the implementation, re-run tests

# If mock setup is incomplete
# → Add proper mock configuration

5. Verify Fix

bash
# Run specific failing test
uv run pytest MouseMaster/Testing/Python/test_file.py::TestClass::test_name -v

# Run full suite
uv run pytest MouseMaster/Testing/Python/ -v

Analyzing Slicer Test Failures

Read Slicer output log

bash
cat ci-artifacts/slicer/slicer-output.log

Common Slicer issues

ErrorCauseFix
ModuleNotFoundErrorExtension not loadedCheck module paths
AttributeError: 'NoneType'Widget not readyAdd processEvents()
RuntimeError: widget deletedView closedCheck object lifetime
Qt assertionInvalid UI operationRun on main thread

Screenshot analysis

If screenshots captured, review with:

bash
# View manifest
cat ci-artifacts/screenshots/manifest.json

# Then use /review-ui-screenshots skill

Decision Rules

Fix immediately

  • •Clear implementation bugs
  • •Missing mock configuration
  • •Import errors
  • •Typos

Investigate more

  • •Intermittent failures
  • •Platform-specific issues
  • •Complex logic errors

Report only

  • •Test infrastructure issues
  • •CI configuration problems
  • •External dependency failures

Report Format

After analysis:

markdown
## Test Analysis Report

### Failures Found
| Test | Error | Root Cause |
|------|-------|------------|
| test_name | AssertionError | Incorrect mock setup |

### Fixes Applied
- `file.py:123` - Fixed mock configuration
- `test.py:45` - Updated assertion

### Remaining Issues
- Issue requiring further investigation

### Verification
- [ ] All tests pass locally
- [ ] Lint passes
- [ ] Type check passes

Integration

After fixing:

  1. •Run /run-tests to verify
  2. •Commit with fix: description of fix
  3. •Push to trigger CI

Related Skills

  • •/download-ci-artifacts - Get artifacts first
  • •/review-ui-screenshots - Analyze UI issues
  • •/fix-bad-practices - Fix code quality issues