command-failing-tests

A Claude Code slash command for debugging and fixing failing tests.

Installation

# Clone to your preferred location
git clone git@github.com:claude-commands/command-failing-tests.git <clone-path>/command-failing-tests

# Symlink (use full path to cloned repo)
ln -s <clone-path>/command-failing-tests/failing-tests.md ~/.claude/commands/failing-tests.md

Usage

/failing-tests              # Run all tests, analyze failures
/failing-tests auth         # Run only auth-related tests
/failing-tests --flaky      # Identify flaky tests
/failing-tests --fix        # Auto-fix simple failures

What it does

1. Runs tests and captures failures
2. Parses error messages and stack traces
3. Analyzes root causes
4. Categorizes by fix type
5. Suggests or applies fixes
6. Verifies fixes pass

Output Format

# Test Failure Analysis

## Summary
- Total: 150 | Passing: 142 | Failing: 8

## Failures
| Test | Error | Root Cause |
|------|-------|------------|
| auth/login | 401 != 200 | Mock not setup |

## Detailed Analysis
### auth/login.test.ts
**Error**: Expected 200, Received 401
**Fix**: Update mock to return valid token

Failure Categories

Category	Example	Fix Type
Assertion	Wrong expected value	Update test
Implementation	Code bug	Fix source
Environment	Missing env var	Fix setup
Flaky	Race condition	Fix timing

Framework Support

Framework	Command	Notes
Jest	npx jest	--json output
Vitest	npx vitest	--reporter=json
Go	go test	-json flag
pytest	pytest	--tb=short

Requirements

• Git repository with tests
• Test framework installed
• Claude Code with Opus 4.5 model access

Updates

cd <clone-path>/command-failing-tests && git pull