Smart Test

Smart Test

Automatically identify and run relevant tests based on your code changes, then fix failures.

When to Use

✅ Use this skill when:

• You've made code changes and need to verify they work
• You want to run only relevant tests (not the full suite)
• Tests are failing and you need to fix them
• You want to improve test coverage

❌ Do NOT use when:

• Writing tests from scratch (use test-driven-development instead)
• Debugging production issues (use systematic-debugging)

Prerequisites

• Test framework installed (Jest, pytest, JUnit, etc.)
• Existing test suite
• Code changes ready to test

Quick Start

# Find tests related to changed files
git diff --name-only | grep -E "\.(js|ts|py|java)$" | while read f; do
  echo "Related test: ${f%.*}.test.${f##*.}"
done

# Run relevant tests
npm test -- --findRelatedTests $(git diff --name-only)

Instructions

Step 1: Identify Changed Files

# What files changed?
git diff --name-only HEAD~1

# Or for staged changes
git diff --cached --name-only

Step 2: Map to Test Files

Common patterns:

# JavaScript/TypeScript
echo "src/components/Button.tsx" | sed 's|src/|tests/|;s|\.tsx$|.test.tsx|'

# Python
echo "lib/utils.py" | sed 's|lib/|tests/test_|;s|\.py$||'

# Java
echo "src/main/User.java" | sed 's|src/main/|src/test/|;s|\.java$|Test.java|'

Step 3: Run Targeted Tests

# Jest (JavaScript)
npx jest --findRelatedTests src/auth/login.js

# pytest (Python)
pytest tests/test_utils.py -v

# Gradle (Java)
./gradlew test --tests "*UserTest*"

# All changed files
git diff --name-only | xargs -I {} sh -c 'echo "Testing: {}" && npm test -- --findRelatedTests {}'

Step 4: Analyze Failures

# Run with verbose output
npm test -- --verbose 2>&1 | tee test-output.log

# Show only failures
grep -A 5 "FAIL" test-output.log

Step 5: Fix and Re-run

# Fix the code or test
vim src/auth/login.js

# Re-run failing test
npm test -- --testPathPattern="login" --watch

Pitfalls

⚠️ Common Mistakes:

1. Running all tests every time

Problem: Full test suite takes too long

Solution: Use targeted test runs:

# Instead of
npm test  # Runs everything

# Do this
npm test -- --findRelatedTests src/changed-file.js

2. Ignoring test output

Problem: Just checking pass/fail, not reading errors

Solution: Read the actual error message:

# Show full error details
npm test -- --verbose 2>&1 | grep -B 5 -A 10 "Error"

3. Fixing tests to pass (not fixing code)

Problem: Changing tests to match broken behavior

Solution:

1. Understand what the test verifies
2. Fix the source code, not the test
3. Only modify test if requirements changed

4. Not testing edge cases

Problem: Tests pass with happy path only

Solution: Add tests for:

• Null/undefined inputs
• Empty arrays/strings
• Boundary values
• Error conditions

Verification

After running tests:

• All tests pass

  npm test 2>&1 | tail -5
  # Should show: Tests: X passed, X total
  

• Coverage is maintained

  npm test -- --coverage
  # Check coverage didn't decrease
  

• No skipped tests

  grep -r "skip\|xit\|xdescribe" tests/
  # Should be empty
  

Examples

Example 1: JavaScript with Jest

# Changed file
git diff --name-only
# src/utils/validation.js

# Run related tests
npx jest --findRelatedTests src/utils/validation.js

# Output:
# PASS tests/utils/validation.test.js
#   validateEmail
#     ✓ should validate correct email
#     ✓ should reject invalid email
#     ✓ should handle empty string
#
# Test Suites: 1 passed, 1 total
# Tests:       3 passed, 3 total

Example 2: Python with pytest

# Changed file
git diff --name-only
# lib/database.py

# Run related tests
pytest tests/test_database.py -v

# Output:
# tests/test_database.py::TestDatabase::test_connection PASSED
# tests/test_database.py::TestDatabase::test_query PASSED
# tests/test_database.py::TestDatabase::test_transaction PASSED

Example 3: Fixing a Failing Test

# Run tests
npm test 2>&1 | grep "FAIL"
# FAIL tests/auth/login.test.js

# See the error
npm test -- --testPathPattern="login" 2>&1 | grep -A 10 "Error"
# Expected: { success: true }
# Received: { success: false, error: "Invalid credentials" }

# Fix the code
vim src/auth/login.js

# Re-run
npm test -- --testPathPattern="login"
# PASS tests/auth/login.test.js

Advanced: Watch Mode

# Jest watch mode - re-runs on file changes
npm test -- --watch

# pytest watch mode (with pytest-watch)
ptw tests/test_utils.py

# Run only failed tests
npm test -- --onlyFailures

References

• Jest Documentation
• pytest Documentation
• Testing Library
• Test Driven Development

Part of Awesome AI Agent Skills