Test Documentation Creator

When to Use This Skill

• Creating only test documentation (testing-strategy.md + tests/README.md)
• Validating existing test documentation structure and content
• Setting up test philosophy and structure documentation for existing project
• NOT creating full documentation structure (use ln-100-documents-pipeline for complete setup)

How It Works

Phase 1: Create Test Documentation

• Check if

  docs/reference/guides/

  exists → create if missing
• Check if

  tests/

  exists → create if missing
• Check if

  tests/manual/

  exists → create if missing
• Check if

  tests/manual/results/

  exists → create if missing
• Add

  tests/manual/results/

  to project

  .gitignore

  if not present
• Log for each: "✓ Created [directory]/" or "✓ [directory]/ already exists"

• Check if

  docs/reference/guides/testing-strategy.md

  exists
• If exists:
  • Skip creation
  • Log: "✓ testing-strategy.md already exists, proceeding to validation"
• If NOT exists:
  • Copy template:

    ln-114-test-docs-creator/references/testing_strategy_template.md

    →

    docs/reference/guides/testing-strategy.md

  • Replace placeholders:

    • [CURRENT_DATE]

      — current date (YYYY-MM-DD)
  • Log: "✓ Created testing-strategy.md from template"
• Check if

  tests/README.md

  exists
• If exists:
  • Skip creation
  • Log: "✓ tests/README.md already exists, proceeding to validation"
• If NOT exists:
  • Copy template:

    ln-114-test-docs-creator/references/tests_readme_template.md

    →

    tests/README.md

  • Replace placeholders:

    • {{DATE}}

      — current date (YYYY-MM-DD)
  • Log: "✓ Created tests/README.md from template"

docs/reference/guides/
└── testing-strategy.md           # Created or existing

tests/
├── README.md                     # Created or existing
└── manual/
    └── results/                  # Created, added to .gitignore

Phase 2: Validate Structure

• Read both files (testing-strategy.md, tests/README.md) - first 5 lines only
• Check for

  <!-- SCOPE: ... -->

  tag in each
• Expected SCOPE tags:
  • testing-strategy.md:

    <!-- SCOPE: Universal testing philosophy (Risk-Based Testing, test pyramid, isolation patterns) -->

  • tests/README.md:

    <!-- SCOPE: Test organization structure (directory layout, Story-Level Test Task Pattern) -->

• If missing in either file:
  • Use Edit tool to add SCOPE tag at line 1 (before first heading)
  • Track violation:

    scope_tags_added += 1

• Load expected sections from

  references/questions.md

• For testing-strategy.md, required sections:
  • "Testing Philosophy"
  • "Test Levels"
• For tests/README.md, required sections:
  • "Test Organization"
  • "Running Tests"
• For each file:
  • Read file content
  • Check if

    ## [Section Name]

    header exists
  • If missing:
    • Use Edit tool to add section with placeholder content from template
    • Track violation:

      missing_sections += 1

• For each file (testing-strategy.md, tests/README.md):
  • Search for

    ## Maintenance

    header
  • If missing:
    • Use Edit tool to add at end of file:
      markdown

      ## Maintenance
      
      **Last Updated:** [current date]
      
      **Update Triggers:**
      - Test framework changes
      - Test organization changes
      - New test patterns introduced
      
      **Verification:**
      - [ ] All test examples follow current framework syntax
      - [ ] Directory structure matches actual tests/
      - [ ] Test runner commands are current

    • Track violation:

      maintenance_added += 1

• For each file:
  • Check if file ends with single blank line (LF)
  • If missing:
    • Use Edit tool to add final newline
    • Track fix:

      posix_fixed += 1

• Log summary:

  ✅ Structure validation complete:
    - SCOPE tags: [added N / already present]
    - Missing sections: [added N sections]
    - Maintenance sections: [added N / already present]
    - POSIX endings: [fixed N / compliant]

• If violations found: "⚠️ Auto-fixed [total] structural violations"

Phase 3: Validate Content

• Read

  references/questions.md

• Parse questions and validation heuristics for 4 sections (2 per file)

1. Testing Philosophy section:
   • Read section content
   • Check validation heuristics from questions.md:
     • ✅ Mentions "Risk-Based Testing"
     • ✅ Has test pyramid description
     • ✅ Mentions priority threshold (≥15)
     • ✅ References Story-Level Test Task Pattern
     • ✅ Length > 100 words
   • If ANY heuristic passes → content valid
   • If ALL fail → log warning: "⚠️ testing-strategy.md → Testing Philosophy section may need review"
2. Test Levels section:
   • Read section content
   • Check validation heuristics from questions.md:
     • ✅ Lists 3 levels (E2E, Integration, Unit)
     • ✅ Has numeric ranges (2-5, 3-8, 5-15)
     • ✅ Explains rationale
     • ✅ Length > 150 words
   • If ANY heuristic passes → content valid
   • If ALL fail → log warning: "⚠️ testing-strategy.md → Test Levels section may need review"

1. Auto-discover test framework:
   • Check

     package.json

     → "devDependencies" and "dependencies":
     • Node.js frameworks: jest, vitest, mocha, ava, tap, jasmine
     • If found: Extract name and version
   • Check

     requirements.txt

     (if Python project):
     • Python frameworks: pytest, nose2, unittest2
     • If found: Extract name and version
   • Check

     go.mod

     (if Go project):
     • Go uses built-in testing package
   • If framework detected:
     • Log: "✓ Test framework detected: [framework]@[version]"
     • Track:

       framework_detected = "[framework]"

   • If NOT detected:
     • Log: "⚠️ No test framework detected. Will use generic test organization."
     • Track:

       framework_detected = None

2. Auto-discover test directory structure:
   • Use Glob tool to scan tests/ directory:
     • Pattern:

       "tests/e2e/**/*.{js,ts,py,go}"

     • Pattern:

       "tests/integration/**/*.{js,ts,py,go}"

     • Pattern:

       "tests/unit/**/*.{js,ts,py,go}"

   • Count files in each directory:

     • e2e_count = len(e2e_files)

     • integration_count = len(integration_files)

     • unit_count = len(unit_files)

   • If directories exist:
     • Log: "✓ Test structure: [e2e_count] E2E, [integration_count] Integration, [unit_count] Unit tests"
   • If directories DON'T exist:
     • Create placeholder structure:

       tests/
         e2e/       (empty, ready for E2E tests)
         integration/  (empty, ready for Integration tests)
         unit/      (empty, ready for Unit tests)

     • Log: "✓ Created test directory structure (will be populated during Story test task execution)"
3. Auto-discover naming conventions:
   • For each test file found (from step 2):
     • Extract filename pattern:

       • *.test.js

         → "*.test.js" convention

       • *.spec.ts

         → "*.spec.ts" convention

       • test_*.py

         → "test_*.py" convention

       • *_test.go

         → "*_test.go" convention
   • Count occurrences of each pattern
   • Use most common pattern (majority rule)
   • If pattern detected:
     • Log: "✓ Naming convention: [pattern] (detected from [count] files)"
   • If NO files exist:
     • Use framework default:
       • Jest/Vitest → *.test.js
       • Mocha → *.spec.js
       • Pytest → test_*.py
       • Go → *_test.go
     • Log: "✓ Naming convention: [default_pattern] (framework default)"
4. Check Test Organization section content:
   • Read section from tests/README.md
   • Check validation heuristics:
     • ✅ Describes directory structure (e2e/integration/unit)
     • ✅ Mentions naming conventions
     • ✅ References Story-Level Test Task Pattern
     • ✅ Has framework mention
   • If ANY heuristic passes → content valid
   • If ALL fail → log warning: "⚠️ tests/README.md → Test Organization section needs update"

1. Auto-discover test runner command:
   • Read

     package.json

     → "scripts" → "test"
   • If found:
     • Extract command value
     • Examples:

       • "jest"

         → Test runner: "npm test" (runs jest)

       • "vitest"

         → Test runner: "npm test" (runs vitest)

       • "mocha"

         → Test runner: "npm test" (runs mocha)
       • Custom script → Test runner: "npm test" (runs [custom])
     • Log: "✓ Test runner: npm test (runs [command])"
   • If NOT found (no package.json or no test script):
     • Use default based on detected framework (from step 3.3.1):
       • Jest → "npm test"
       • Vitest → "npm test"
       • Pytest → "pytest"
       • Go → "go test ./..."
     • Log: "⚠️ No test script found in package.json. Using default '[command]'."
2. Auto-discover coverage command (optional):
   • Check

     package.json

     → "scripts" for:
     • "test:coverage"
     • "coverage"
     • "test:cov"
   • If found:
     • Extract command
     • Log: "✓ Coverage command: npm run [script_name]"
   • If NOT found:
     • Use framework default:
       • Jest → "npm test -- --coverage"
       • Vitest → "npm test -- --coverage"
       • Pytest → "pytest --cov=src"
       • Go → "go test -cover ./..."
     • Log: "✓ Coverage command: [default] (framework default)"
3. Check Running Tests section content:
   • Read section from tests/README.md
   • Check validation heuristics:
     • ✅ Has test runner command
     • ✅ Mentions coverage
     • ✅ Shows how to run specific tests
     • ✅ Has command examples
   • If ANY heuristic passes → content valid
   • If ALL fail → log warning: "⚠️ tests/README.md → Running Tests section needs update"

• Log summary:

  ✅ Content validation complete:
    - testing-strategy.md: [2 sections checked]
    - tests/README.md: [2 sections checked]
    - Test framework: [detected framework or "Not detected"]
    - Test structure: [e2e/integration/unit counts or "Created placeholder"]
    - Naming convention: [pattern or "Framework default"]
    - Test runner: [command]
    - Coverage command: [command]

Complete Output Structure

docs/reference/guides/
└── testing-strategy.md           # Universal testing philosophy (465 lines)

tests/
└── README.md                     # Test organization + Story-Level Pattern (112 lines)

Reference Files Used

• Risk-based testing methodology:

  shared/references/risk_based_testing.md

Templates

• references/testing_strategy_template.md

  - Universal testing philosophy with:
  • SCOPE tags (testing philosophy, NOT framework-specific)
  • Core Philosophy ("Test YOUR code, not frameworks")
  • Risk-Based Testing Strategy (Priority Matrix, test caps)
  • Story-Level Testing Pattern
  • Test Organization (E2E/Integration/Unit definitions)
  • Isolation Patterns (Data Deletion/Transaction Rollback/DB Recreation)
  • What To Test vs NOT Test (universal examples)
  • Testing Patterns (Arrange-Act-Assert, Mock at the Seam, Test Data Builders)
  • Common Issues (Flaky Tests, Slow Tests, Test Coupling)
  • Coverage Guidelines
  • Verification Checklist

• references/tests_readme_template.md

  - Test organization with:
  • SCOPE tags (test documentation ONLY)
  • Overview (E2E/Integration/Unit test organization)
  • Testing Philosophy (brief, link to testing-strategy.md)
  • Test Structure (directory tree)
  • Story-Level Test Task Pattern (tests in final Story task, NOT scattered)
  • Test Execution (project-specific commands)
  • Quick Navigation (links to testing-strategy.md, kanban_board, guidelines)
  • Maintenance section (Update Triggers, Verification, Last Updated)

• references/questions.md

  (v1.0.0) - Question-driven validation:
  • Questions each section must answer (4 sections total)
  • Validation heuristics (ANY passes → valid)
  • Auto-discovery hints (test frameworks, directory structure, naming conventions)
  • MCP Ref hints (external research if needed)

Best Practices

• No premature validation: Phase 1 creates structure, Phase 2 validates it (no duplicate checks)
• Parametric validation: Phase 3 validates 4 sections across 2 files (no code duplication)
• Auto-discovery first: Scan test frameworks and directory structure before using defaults
• Idempotent: ✅ Can run multiple times safely (checks existence before creation, re-validates on each run)
• Separation of concerns: CREATE → VALIDATE STRUCTURE → VALIDATE CONTENT
• Story-Level Test Task Pattern: Tests consolidated in final Story task (ln-510-test-planner creates task, ln-420-test-executor implements)
• Value-Based Testing: 2-5 E2E, 3-8 Integration, 5-15 Unit per Story (10-28 total max), Priority ≥15 MUST be tested
• No test code: This skill creates DOCUMENTATION only, NOT actual tests

Documentation Standards

• NO_CODE Rule: Test docs describe strategy, not test implementations
• Stack Adaptation: Framework commands must match project stack
• Format Priority: Tables (test levels, counts) > Lists > Text

Prerequisites

• docs/reference/guides/

  directory (created by ln-112-reference-docs-creator or Phase 1 if missing)

• docs/reference/guides/testing-strategy.md

  (universal testing philosophy)

• tests/README.md

  (test organization structure)
• Validated structure and content (auto-discovery of test frameworks and directory structure)

Definition of Done

• docs/reference/guides/

  directory exists

• tests/

  directory exists

• tests/manual/

  directory exists

• tests/manual/results/

  directory exists

• tests/manual/results/

  added to

  .gitignore

• testing-strategy.md

  exists (created or existing)

• tests/README.md

  exists (created or existing)

• SCOPE tags present in both files (first 5 lines)
• testing-strategy.md has "Testing Philosophy" and "Test Levels" sections
• tests/README.md has "Test Organization" and "Running Tests" sections
• Maintenance sections present in both files at end
• POSIX file endings compliant (LF, single blank line at EOF)

• testing-strategy.md → Testing Philosophy section checked (Risk-Based Testing mentioned)
• testing-strategy.md → Test Levels section checked (2-5 E2E, 3-8 Integration, 5-15 Unit)
• tests/README.md → Test Organization section checked or auto-discovered
• tests/README.md → Running Tests section checked or auto-discovered
• Test framework detected (if applicable) and logged
• Test directory structure scanned or created
• Naming conventions detected or defaults used
• Test runner command identified or defaults used

• Phase 1 logged: creation summary
• Phase 2 logged: structural fixes (if any)
• Phase 3 logged: content validation summary with auto-discovery results

Technical Details

• Story-Level Test Task Pattern
• Value-Based Testing (2-5 E2E, 3-8 Integration, 5-15 Unit, 10-28 total max per Story)
• Risk-Based Testing (Priority ≥15)

• Node.js: jest, vitest, mocha, ava, tap, jasmine
• Python: pytest, nose2, unittest2
• Go: built-in testing package