ln-140-test-docs-creator（ln 140 测试文档创建者）

Paths: File paths (shared/, references/, ../ln-*) are relative to skills repo root. If not found at CWD, locate this SKILL.md directory and go up one level for repo root. If shared/ is missing, fetch files via WebFetch from https://raw.githubusercontent.com/levnikolaevich/claude-code-skills/master/skills/{path}.

Test Documentation Creator

Type: L2 Worker

This skill creates and validates test documentation: testing-strategy.md (universal testing philosophy) + tests/README.md (test organization structure with tests/automated/ + Story-Level Test Task Pattern).

Purpose

Creates and validates test documentation (testing-strategy.md + tests/README.md) establishing universal testing philosophy, Risk-Based Testing strategy, tests/automated/ as the default automated-test root, and the Story-Level Test Task Pattern for any project.

When to Use This Skill

This skill is a L2 WORKER invoked by ln-100-documents-pipeline orchestrator.

This skill should be used directly when:

• 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)

Workflow

The skill follows a 3-phase workflow: CREATE → VALIDATE STRUCTURE → VALIDATE CONTENT. Each phase builds on the previous, ensuring complete structure and semantic validation.

MANDATORY READ: Load shared/references/docs_quality_contract.md, shared/references/docs_quality_rules.json, and shared/references/markdown_read_protocol.md.

Phase 1: Create Test Documentation

Objective: Establish test philosophy and documentation structure.

Process:

1.1 Check & create directories:

• 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"

1.2 Check & create documentation files:

• 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-140-test-docs-creator/references/testing_strategy_template.md → docs/reference/guides/testing-strategy.md - Replace placeholders: - [CURRENT_DATE] — current date (YYYY-MM-DD) - Preserve shared metadata markers and standard top sections from the template - 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-140-test-docs-creator/references/tests_readme_template.md → tests/README.md - Replace placeholders: - {{DATE}} — current date (YYYY-MM-DD) - Preserve shared metadata markers and standard top sections from the template - Log: "✓ Created tests/README.md from template"

1.3 Output:

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

Objective: Ensure test documentation files comply with structural requirements and auto-fix violations.

Process:

2.1 Check SCOPE tags:

• Read both files (testing-strategy.md, tests/README.md) - opening block first
• Check for <!-- SCOPE:... --> tag and metadata markers 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

2.2 Check required sections:

• 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 the section with minimal concrete guidance from the template, not a raw placeholder - Track violation: missing_sections += 1

2.3 Check Maintenance section:

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

- Search for ## Maintenance header - If missing: - Use Edit tool to add at end of file: ## 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

2.4 Check POSIX file endings:

• 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

2.5 Report validation:

• 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

Objective: Ensure each section answers its questions with meaningful content and populate test-specific sections from auto-discovery.

Process:

3.1 Load validation spec:

• Read references/questions.md
• Parse questions and validation heuristics for 4 sections (2 per file)

3.2 Validate testing-strategy.md sections:

For this file, use standard template content (no auto-discovery needed):

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"

1. 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"

Note: testing-strategy.md is universal philosophy - no project-specific auto-discovery needed.

3.3 Validate tests/README.md sections with auto-discovery:

Section: Test Organization

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

1. Auto-discover test directory structure:

- Use Glob tool to scan tests/ directory: - Pattern: "tests/automated/e2e/**/*.{js,ts,py,go}" - Pattern: "tests/automated/integration/**/*.{js,ts,py,go}" - Pattern: "tests/automated/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/ automated/ 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)"

1. 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)"

1. Check Test Organization section content:

- Read section from tests/README.md - Check validation heuristics: - ✅ Describes directory structure (tests/automated/e2e, tests/automated/integration, tests/automated/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"

Section: Running Tests

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]'."

1. 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)"

1. 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"

3.4 Report content validation:

• 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: [automated/e2e, automated/integration, automated/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)

Note: Actual test directories (e2e/, integration/, unit/) created during Story test task execution or Phase 3 if missing.

Critical Rules

• Documentation only: This skill creates test DOCUMENTATION, NOT actual test code
• 3-phase pipeline: CREATE → VALIDATE STRUCTURE → VALIDATE CONTENT (no phase skipping)
• Auto-discovery first: Scan test frameworks and directory structure before falling back to defaults
• Idempotent execution: Checks existence before creation; re-validates on each run without duplication
• Shared opening contract required: Both files must include SCOPE, metadata markers, Quick Navigation, Agent Entry, and Maintenance

Reference Files

• Risk-based testing methodology: shared/references/risk_based_testing_guide.md

Templates

Testing Strategy Template:

• 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 (tests/automated/e2e, tests/automated/integration, tests/automated/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

Tests README Template:

• 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)

Validation Specification:

• 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 (test planner creates task, test executor implements)
• Value-Based Testing: Priority ≥15 MUST be tested, each test justified by Usefulness Criteria
• 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
• Shared docs-quality contract: SCOPE tags, Maintenance sections, placeholder policy, and stack-specific link validation come from the shared docs-quality contract/rules

Prerequisites

Invoked by: ln-100-documents-pipeline orchestrator

Requires:

• docs/reference/guides/ directory (created by ln-120-reference-docs-creator or Phase 1 if missing)

Creates:

• 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)

Return Contract

Return a normalized summary so ln-100 can run one centralized docs-quality gate:

{
  "created_files": [
    "docs/reference/guides/testing-strategy.md",
    "tests/README.md"
  ],
  "skipped_files": [],
  "quality_inputs": {
    "doc_paths": [
      "docs/reference/guides/testing-strategy.md",
      "tests/README.md"
    ],
    "owners": {
      "docs/reference/guides/testing-strategy.md": "ln-140-test-docs-creator",
      "tests/README.md": "ln-140-test-docs-creator"
    }
  },
  "validation_status": "passed|passed_with_fixes|skipped"
}

Runtime Summary Artifact

MANDATORY READ: Load shared/references/docs_generation_summary_contract.md

Accept optional summaryArtifactPath.

Summary kind:

• docs-generation

Required payload semantics:

• worker = "ln-140"
• status
• created_files
• skipped_files
• quality_inputs
• validation_status
• warnings

Write the summary to the provided artifact path or return the same envelope in structured output.

Definition of Done

Before completing work, verify ALL checkpoints:

✅ Structure Created (Phase 1):

• 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)

✅ Structure Validated (Phase 2):

• 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)

✅ Content Validated (Phase 3):

• testing-strategy.md → Testing Philosophy section checked (Risk-Based Testing mentioned)
• testing-strategy.md → Test Usefulness Criteria section checked
• 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

✅ Reporting:

• Phase 1 logged: creation summary
• Phase 2 logged: structural fixes (if any)
• Phase 3 logged: content validation summary with auto-discovery results
• Return contract emitted with created_files, skipped_files, quality_inputs, and validation_status

Technical Details

Standards:

• Story-Level Test Task Pattern
• Risk-Based Testing (Priority ≥15, Usefulness Criteria)

Language: English only

Auto-Discovery Support:

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

Version: 7.2.0 Last Updated: 2026-01-15