Test Suite Optimization

Autonomous analysis to identify slow and redundant tests, create optimization issues.

Usage

/test-optimize              # Analyze all modules
/test-optimize src/auth     # Analyze specific module
/test-optimize --dry-run    # Report only, don't create issues

Instructions

1. Bootstrap State (if needed)

Check for .testaudit.yaml:

If missing:

• Create with template
• Find all CLAUDE.md directories (these are modules)
• Initialize each module with current HEAD commit

version: 1

modules:
  src/auth:
    test_count: 0
    total_duration_ms: 0
    avg_duration_ms: 0
    slow_tests: 0
    last_analyzed_commit: <current-HEAD>
    last_analyzed_date: <now-ISO8601>

2. Module Discovery

Find all directories with CLAUDE.md - these define modules for analysis.

If no CLAUDE.md files found:

Error: No modules found (no CLAUDE.md files)
Run '/docsaudit' first to create CLAUDE.md hierarchy.

3. Determine Modules to Analyze

• If specific module provided: analyze only that module
• Otherwise: analyze all modules with tests

4. Run Analysis

Use test-optimization skill for domain knowledge on:

• Speed analysis (slow test causes and solutions)
• Redundancy detection (semantic overlap, logical subsumption)
• DRY analysis (fixtures, parameterization, helpers)

Run slowtests to identify bottlenecks:

just slowtests 50

If unavailable, run tests directly with timing output.

5. Prioritize Opportunities

Use test-optimization skill prioritization formula:

• Impact score (speed improvement, maintainability, test reduction)
• Complexity score (mocking vs redesign)
• Priority = impact / (complexity + 1)

6. Check Existing Issues

gh issue list --search "is:open [test-optimization]" --json number,title,body

Skip opportunities with similar existing issues.

7. Create GitHub Issues

For each new opportunity:

[test-optimization] <Short description>

## Current State
- Test file: <file>
- Slow tests: <count> (avg: <ms>ms, target: <50ms)
- Redundant tests: <count>
- DRY opportunities: <count>

## Issues
[Speed, redundancy, and DRY issues identified]

## Approach
[Prioritized fix steps]

## Expected Improvement
- Speed: <current>ms -> <target>ms
- Maintainability: <tests removed/consolidated>

## Acceptance Criteria
- [ ] All tests run in <50ms each
- [ ] Coverage maintained at 96%
- [ ] just check-all passes

8. Update State File

Update .testaudit.yaml with current analysis results.

9. Generate Report

Test Suite Optimization Report
==============================

Analyzed modules: <N>

Module Test Health:
  <module>: <tests> tests, <duration>s total, <avg>ms avg

Opportunities found: <N>
Issues created: <list>

Next: /work <issue-number>

Options

• --dry-run: Report only, don't create issues or update state
• <module-path>: Analyze only specified module

Error Handling

• No CLAUDE.md files: Suggest /docsaudit
• No just slowtests: Use direct test execution
• GitHub CLI missing: Suggest --dry-run

Notes

• Requires CLAUDE.md files (run /docsaudit first)
• Works best with justfile Level 1 (just slowtests)
• Target: all unit tests <50ms each
• Safe to run repeatedly (checks for changes, avoids duplicates)