Skip to main content Skills Marketplace Discover and explore AI skills built by the community.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Show prompt details A direct command skips the review prompt. Inspect the source before running it.
npx skills add https://github.com/diegosouzapw/awesome-omni-skill --skill ln-100-documents-pipelineThe command stays on one line. Scroll horizontally to inspect it before copying.
Prefer a local copy? Download the files currently available to SkillsMP.
Download Zip Downloading... More from this repository Token-efficient tracking for AI orchestration. CLI-first for status updates (~50 tokens), agent fallback for complex ops (~1KB). Use when: updating task status, querying blockers, creating progress files, validating phases.
AshAi extension guidelines for integrating AI capabilities with Ash Framework. Use when implementing vectorization/embeddings, exposing Ash actions as LLM tools, creating prompt-backed actions, or setting up MCP servers. Covers semantic search, LangChain integration, and structured outputs.
This skill should be used when solving hard questions, complex architectural problems, or debugging issues that benefit from GPT-5 Pro or GPT-5.1 thinking models with large file context. Use when standard Claude analysis needs deeper reasoning or extended context windows.
Related occupations SOC
Based on SOC occupation classification
name ln-100-documents-pipeline description Top orchestrator for complete doc system. Delegates to ln-110 coordinator (project docs via 5 L3 workers) + ln-120-150 workers. Phase 4: global cleanup. Idempotent.
Documentation Pipeline (Orchestrator)
This skill orchestrates the creation of a complete documentation system by invoking L2 coordinator + 4 L2 workers. The coordinator (ln-110) delegates to 5 L3 workers for project docs; other L2 workers handle reference/tasks/test/presentation domains. Each component validates its own output.
Architecture
ln-100-documents-pipeline (L1 Top Orchestrator - this skill)
├── ln-110-project-docs-coordinator (L2 Coordinator)
│ ├── ln-111-root-docs-creator (L3 Worker) → 4 docs
│ ├── ln-112-project-core-creator (L3 Worker) → 3 docs
│ ├── ln-113-backend-docs-creator (L3 Worker) → 2 conditional
│ ├── ln-114-frontend-docs-creator (L3 Worker) → 1 conditional
│ └── ln-115-devops-docs-creator (L3 Worker) → 1 conditional
├── ln-120-reference-docs-creator (L2 Worker)
├── ln-130-tasks-docs-creator (L2 Worker)
├── ln-140-test-docs-creator (L2 Worker - optional)
├── ln-150-presentation-creator (L2 Worker)
├── ln-600-docs-auditor (L2 Worker - optional)
└── ln-610-code-comments-auditor (L2 Worker - optional)
When to Use This Skill
This skill should be used when:
Start a new IT project and need complete documentation system at once
Use automated workflow instead of manually invoking multiple workers
Create entire documentation structure (CLAUDE.md → docs/ → presentation/) in one go
Prefer orchestrated CREATE path over manual skill chaining
Need automatic global cleanup (deduplication, orphaned files, consolidation)
Alternative : If you prefer granular control, invoke workers manually:
ln-110-project-docs-coordinator - Root + Project docs (coordinates 5 L3 workers)ln-120-reference-docs-creator - reference/ structure
Note : Each worker now validates its own output in Phase 2/3. Orchestrator handles global operations only.
How It Works The skill follows a 7-phase orchestration workflow: Legacy Migration (optional) → User confirmation → Invoke coordinator + 4 workers sequentially → Global cleanup → Documentation Audit (optional) → Summary. Phase 3 (validation) is intentionally skipped - each component validates its own output.
Phase 0: Legacy Migration (OPTIONAL) Objective : Detect existing documentation in non-standard formats, extract valuable content, and prepare for migration.
Trigger : Always runs at pipeline start. User can skip if no legacy docs or wants to keep existing structure.
Scan project for non-standard documentation using patterns from references/legacy_detection_patterns.md:
Root .md files : ARCHITECTURE.md, REQUIREMENTS.md, STACK.md, API.md, DATABASE.md, DEPLOYMENT.mdLegacy folders : documentation/, doc/, wiki/, docs/ with wrong structureREADME.md sections : ## Architecture, ## Tech Stack, ## Requirements, etc.CONTRIBUTING.md sections : ## Development, ## Code Style, ## Coding Standards
Build legacy_manifest: list of { path, detected_type, target_doc, confidence }
If no legacy docs found → skip to Phase 1
For each detected legacy file:
Parse markdown structure (headers, lists, code blocks)
Apply type-specific extractor (see legacy_detection_patterns.md):
architecture_extractor → { layers[], components[], diagrams[] }requirements_extractor → { functional[], non_functional[] }tech_stack_extractor → { frontend, backend, database, versions }principles_extractor → { principles[], anti_patterns[] }api_spec_extractor → { endpoints[], authentication }database_schema_extractor → { tables[], relationships[] }runbook_extractor → { prerequisites[], install_steps[], env_vars[] }
Score content quality (0.0-1.0)
Store in extracted_content object
Build migration_context from extracted content:
{
"LEGACY_CONTENT" : {
"legacy_architecture" : { "sections" : [ ...] , "diagrams" : [ ...] } ,
"legacy_requirements" : { "functional" : [ ...] } ,
"legacy_principles" : { "principles" : [ ...] } ,
"legacy_tech_stack" : { "frontend" : "..." , "backend" : "..." } ,
"legacy_api" : { "endpoints" : [ ...] } ,
"legacy_database" : { "tables" : [ ...] } ,
"legacy_runbook" : { "install_steps" : [ ...] }
}
}
Merge into Context Store for ln-110:
contextStore.LEGACY_CONTENT = migration_contextWorkers use LEGACY_CONTENT as base content (priority over template defaults)
Priority order: Legacy content > Auto-discovery > Template defaults
0.6 Cleanup (Legacy Files) :
For root-level files (README.md, CONTRIBUTING.md):
Do NOT delete
Remove migrated sections using Edit tool
Add links to new locations:
## Architecture → See [Architecture](docs/project/architecture.md)## Tech Stack → See [Tech Stack](docs/project/tech_stack.md)
For standalone legacy files (ARCHITECTURE.md, documentation/):
Delete files (already backed up)
Log: "Deleted: ARCHITECTURE.md (migrated to docs/project/architecture.md)"
Clean empty legacy directories
Output : migration_summary { migrated_count, archived_count, skipped_count, legacy_content }
Phase 1: User Confirmation Objective : Check existing files, explain workflow, and get user approval.
Migration Summary (if Phase 0 ran):
Show: "✓ Migrated {N} legacy documents"
Show: "✓ Archived to .archive/legacy-{date}/"
Show: "✓ LEGACY_CONTENT prepared for workers"
Pre-flight Check (scan existing documentation):
Show user what will be created:
Root + Project documentation (CLAUDE.md + docs/README.md + documentation_standards.md + principles.md + docs/project/ via ln-110-project-docs-coordinator)
Reference structure (docs/reference/ via ln-120-reference-docs-creator)
Task management docs (docs/tasks/ via ln-130-tasks-docs-creator)
Test documentation (tests/ via ln-140-test-docs-creator - optional)
HTML presentation (docs/presentation/ via ln-150-presentation-creator)
Estimated time: 15-20 minutes with interactive dialog
Ask: "Proceed with creating missing files? (yes/no)"
Ask: "Include test documentation (tests/README.md)? (yes/no)"
Output : File scan summary + user approval + test docs preference
Phase 2: Invoke Coordinator + Workers Sequentially Objective : Create complete documentation system by invoking L2 coordinator + 4 L2 workers in order.
Process (AUTOMATIC invocations with Skill tool):
2.1 Create Root + Project Documentation :
Invocation : Skill(skill: "ln-110-project-docs-coordinator") → AUTOMATICInput : Pass LEGACY_CONTENT from Phase 0 (if migration was performed)Behavior : Coordinator gathers context ONCE, then delegates to 5 L3 workers:
ln-111-root-docs-creator → 4 root docs (uses LEGACY_CONTENT.legacy_principles if available)
ln-112-project-core-creator → 3 core docs (uses LEGACY_CONTENT.legacy_architecture, legacy_requirements, legacy_tech_stack)
ln-113-backend-docs-creator → 2 conditional (uses LEGACY_CONTENT.legacy_api, legacy_database)
ln-114-frontend-docs-creator → 1 conditional (if hasFrontend)
ln-115-devops-docs-creator → 1 conditional (uses LEGACY_CONTENT.legacy_runbook)
Output : Root docs (CLAUDE.md + docs/README.md + docs/documentation_standards.md + docs/principles.md) + Project docs (docs/project/requirements.md, architecture.md, tech_stack.md + conditional: api_spec.md, database_schema.md, design_guidelines.md, runbook.md)Store : Save context_store from ln-110 result (contains TECH_STACK for ln-120)Validation : Each L3 worker validates output (SCOPE tags, Maintenance sections)Verify : All documents exist before continuing 2.2 Create Reference Structure + Smart Documents :
Invocation : Skill(skill: "ln-120-reference-docs-creator") → AUTOMATICInput : Pass context_store from ln-110 (TECH_STACK enables smart document creation)Output : docs/reference/README.md + adrs/, guides/, manuals/, research/ directories + justified ADRs/Guides/Manuals Smart Creation : Creates documents only for nontrivial technology choices (see ln-120 justification rules)Validation : ln-120 validates output in Phase 2/3Verify : Reference hub exists before continuing 2.3 Create Task Management Docs :
Invocation : Skill(skill: "ln-130-tasks-docs-creator") → AUTOMATICOutput : docs/tasks/README.md + optionally kanban_board.md (if user provides Linear config)Validation : ln-130 validates output in Phase 2/3Verify : Tasks README exists before continuing 2.4 Create Test Documentation (Optional) :
Condition : If user approved test docs in Phase 1Invocation : Skill(skill: "ln-140-test-docs-creator") → AUTOMATICOutput : tests/README.md (test documentation with Story-Level Test Task Pattern)Validation : ln-140 validates output in Phase 2/3Skip : If "no" → can run ln-140-test-docs-creator later manually 2.5 Create HTML Presentation :
Invocation : Skill(skill: "ln-150-presentation-creator") → AUTOMATICOutput : docs/presentation/README.md + presentation_final.html + assets/Validation : ln-150 validates output in Phase 2/3Verify : Presentation files exist before continuing Output : Complete documentation system with coordinator + 4 workers completed and validated
TodoWrite format (mandatory):
Add ALL invocations to todos before starting:
- Invoke ln-110-project-docs-coordinator (pending)
- Invoke ln-120-reference-docs-creator (pending)
- Invoke ln-130-tasks-docs-creator (pending)
- Invoke ln-140-test-docs-creator (pending)
- Invoke ln-150-presentation-creator (pending)
- Run Global Cleanup (Phase 4) (pending)
- Run Documentation Audit (Phase 5 - optional) (pending)
Mark each as in_progress when starting, completed when worker returns success.
Phase 4: Global Cleanup and Consolidation Objective : Remove duplicates, orphaned files, consolidate knowledge across ALL documentation.
Trigger : Only after ALL workers complete Phase 2/3 validation.
4.1 Scan for duplicate content
Read all .md files in docs/
Use Glob tool: pattern: "docs/**/*.md"
Store list of all documentation files
Identify duplicate sections:
For each file:
Read file content
Extract section headers (##, ###)
Extract section content (text between headers)
Compare sections across files:
If 2+ sections have:
Same heading (case-insensitive)
80% content similarity (simple word overlap check)
Mark as duplicate
Determine canonical version:
Rules for canonical location:
"Development Principles" → principles.md (always canonical)
"Testing Strategy" → testing-strategy.md (always canonical)
"Linear Configuration" → tasks/kanban_board.md (always canonical)
For other duplicates: Keep in FIRST file encountered (alphabetical order)
Remove duplicates:
For each duplicate section:
Track count of removed duplicates
Log results:
"✓ Removed {count} duplicate sections"
List: "{section_name} removed from {file} (canonical: {canonical_file})"
4.2 Scan for orphaned files
List all .md files in docs/
Use Glob tool: pattern: "docs/**/*.md"
Check against expected structure:
Expected files (created by workers):
docs/CLAUDE.md
docs/README.md
docs/documentation_standards.md
docs/principles.md
docs/project/requirements.md
docs/project/architecture.md
docs/project/tech_stack.md
docs/project/api_spec.md (conditional)
docs/project/database_schema.md (conditional)
docs/project/design_guidelines.md (conditional)
docs/project/runbook.md (conditional)
docs/reference/README.md
docs/reference/adrs/*.md (user-created)
docs/reference/guides/*.md (user-created)
docs/reference/manuals/*.md (user-created)
docs/tasks/README.md
docs/tasks/kanban_board.md
docs/testing-strategy.md
tests/README.md
Any file NOT in this list = orphaned
Move orphaned files to archive:
Create .archive/YYYY-MM-DD/ directory (current date)
For each orphaned file:
Use Bash tool: mv {file_path} .archive/YYYY-MM-DD/
Log: "Archived {file_name} (not in expected structure)"
Track count
Log results:
"✓ Archived {count} orphaned files to .archive/{date}/"
List archived files
4.3 Consolidate knowledge
Identify scattered information:
Known patterns:
Linear config scattered: kanban_board.md + tasks/README.md
Development principles scattered: principles.md + architecture.md + CLAUDE.md
Testing info scattered: testing-strategy.md + tests/README.md + architecture.md
For each scattered concept:
Determine Single Source of Truth (SSoT):
Linear config → kanban_board.md
Development principles → principles.md
Testing strategy → testing-strategy.md
Update non-SSoT files:
Log results:
"✓ Consolidated {count} scattered concepts"
List: "{concept} consolidated to {SSoT_file}"
4.4 Cross-link validation
Scan all .md files for internal links:
For each file:
Read content
Extract all markdown links: [text](path)
Filter internal links (relative paths, not http://)
Verify link targets exist:
For each link:
Resolve relative path
Check if target file exists (Glob tool)
If missing: Mark as broken
Fix broken links:
For each broken link:
If target was archived: Update link to archive path
If target never existed: Remove link (convert to plain text)
Track fix count
Add missing critical links:
CLAUDE.md → docs/README.md:
Read CLAUDE.md
Check for link to docs/README.md
If missing: Add in "Documentation Hub" section
docs/README.md → section READMEs:
Check for links to project/, reference/, tasks/, tests/ READMEs
If missing: Add
Track added links count
Log results:
"✓ Fixed {count} broken links"
"✓ Added {count} missing critical links"
List changes
✅ Global Cleanup Complete:
Structure:
- Removed {N} duplicate sections (canonical: principles.md)
- Archived {N} orphaned files to .archive/YYYY-MM-DD/
- list of archived files
- Consolidated {N} scattered concepts
Links:
- Fixed {N} broken links
- Added {N} missing critical links:
- list of added links
Output : All documentation cleaned up, duplicates removed, orphaned files archived, knowledge consolidated, cross-links validated
Phase 5: Documentation Audit (OPTIONAL) Objective : Audit documentation and code comments quality.
Trigger : Only if user requests audit OR pipeline invoked with audit flags.
📊 Documentation Audit Options:
1. AUDIT DOCS: Run ln-600-docs-auditor (6 categories)
2. AUDIT COMMENTS: Run ln-610-code-comments-auditor (6 categories)
3. BOTH: Run both auditors
4. SKIP: Continue to summary
Choose option (1/2/3/4): _
5.2 Run Selected Auditors :
If AUDIT DOCS selected:
Invocation : Skill(skill: "ln-600-docs-auditor") → AUTOMATICOutput : Compliance Score X/10 per category + Findings
If AUDIT COMMENTS selected:
Invocation : Skill(skill: "ln-610-code-comments-auditor") → AUTOMATICOutput : Compliance Score X/10 per category + Findings
📊 Audit Results:
- Documentation Quality: X/10 overall
- Hierarchy & Links: X/10
- Single Source of Truth: X/10
- ...
- Code Comments Quality: X/10 overall
- WHY not WHAT: X/10
- Density (15-20%): X/10
- ...
See full reports above for detailed findings.
Output : Audit reports with compliance scores and findings
Phase 6: Summary and Next Steps Objective : Provide complete overview of created system.
List all created files with sizes:
CLAUDE.md (project entry point)docs/README.md (root documentation hub)docs/documentation_standards.md (60 universal requirements)docs/principles.md (11 development principles)docs/project/requirements.md, architecture.md, tech_stack.md + conditional documents (3-7 total)docs/reference/README.md (reference hub with empty adrs/, guides/, manuals/, research/ directories)docs/tasks/README.md + optionally kanban_board.mddocs/presentation/README.md + presentation_final.htmltests/README.md (if created)
Show documentation system features:
✅ SCOPE tags (document boundaries defined)
✅ Maintenance sections (update triggers + verification)
✅ README hub (central navigation)
✅ DAG structure (Directed Acyclic Graph navigation)
✅ Living documentation ready
✅ Deduplicated content (canonical sources only)
✅ Validated cross-links (no broken links)
Recommend next steps:
"Review generated documentation (CLAUDE.md → docs/)"
"Open docs/presentation/presentation_final.html in browser"
"Run ln-210-epic-coordinator to decompose scope into Epics"
"Share documentation with technical stakeholders"
Output : Summary message with file list and recommendations
Complete Output Structure project_root/
├── CLAUDE.md # ← Project entry point (link to docs/)
├── docs/
│ ├── README.md # ← Root documentation hub (general standards)
│ ├── documentation_standards.md # ← 60 universal requirements (Claude Code + industry standards)
│ ├── principles.md # ← 11 development principles (Standards First, YAGNI, KISS, DRY, etc.)
│ ├── project/
│ │ ├── requirements.md # ← Functional Requirements (NO NFR per project policy)
│ │ ├── architecture.md # ← arc42-based architecture with C4 Model
│ │ ├── tech_stack.md # ← Technology versions, Docker config
│ │ ├── api_spec.md # ← API endpoints (conditional)
│ │ ├── database_schema.md # ← Database schema (conditional)
│ │ ├── design_guidelines.md # ← UI/UX system (conditional)
│ │ └── runbook.md # ← Operations guide (conditional)
│ ├── reference/
│ │ ├── README.md # ← Reference documentation hub (registries)
│ │ ├── adrs/ # ← Empty, ready for ADRs
│ │ ├── guides/ # ← Empty, ready for guides
│ │ ├── manuals/ # ← Empty, ready for manuals
│ │ └── research/ # ← Empty, ready for research
│ ├── tasks/
│ │ ├── README.md # ← Task management system rules
│ │ └── kanban_board.md # ← Linear integration (optional)
│ └── presentation/
│ ├── README.md # ← Navigation hub for presentation
│ ├── presentation_final.html # ← Final standalone HTML (~130-180 KB)
│ └── assets/ # ← Modular HTML structure
└── tests/
└── README.md # ← Test documentation (optional)
Integration with Project Workflow Recommended workflow for new projects:
ln-100-documents-pipeline (this skill) - Create complete documentation systemln-210-epic-coordinator - Decompose scope into Epics (Linear Projects)ln-220-story-coordinator - Create User Stories for each Epic (automatic decomposition + replan)ln-300-task-coordinator - Break down Stories into implementation tasks (automatic decomposition + replan)ln-310-story-validator - Verify Stories before developmentln-400-story-executor - Orchestrate Story implementationln-500-story-quality-gate - Review completed Stories
Advantages of Orchestrator Approach
✅ Single command creates complete system
✅ No need to remember skill sequence
✅ Automated skill chaining
✅ Consistent output every time
✅ Time-saving (one invocation vs 2-3 manual invocations)
✅ Idempotent : Safe to run multiple times - preserves existing files, creates only missing ones
✅ Pre-flight check : Shows file scan summary before execution
✅ Global cleanup : Automatic deduplication, orphaned files archival, knowledge consolidation
✅ Validated cross-links : No broken links in documentation
⚠️ Less granular control (can't skip coordinator phases)
⚠️ Longer execution time (15-20 minutes)
⚠️ Global cleanup runs even if only one file was created
When to use manual approach instead:
Documentation Standards All documents created by this pipeline MUST follow these rules:
Rule Description Enforcement NO_CODE Rule Documents describe contracts, not implementations No code blocks >5 lines; use tables/ASCII/links Stack Adaptation Links must match project TECH_STACK .NET → Microsoft docs, JS → MDN Format Priority Tables/ASCII > Lists (enumerations only) > Text Tables for params, config, alternatives
These standards are enforced by L3 workers (ln-111-115) and audited by ln-600-docs-auditor.
Error Handling If any invoked skill fails:
Notify user which skill failed
Show error message from failed skill
Recommend manual invocation for debugging
List already completed steps (partial progress)
Technical Implementation Notes
Uses Skill tool with command parameter
Waits for each skill to complete before proceeding
Verifies output files exist before moving to next phase
Uses Glob tool to check docs/project/ structure
Lists file sizes for user confirmation
Warns if expected files missing
Uses Glob tool to find all .md files
Uses Read tool to analyze content
Uses Edit tool to remove duplicates and add links
Uses Bash tool to archive orphaned files
All output follows same standards as underlying skills
ISO/IEC/IEEE 29148:2018 (Requirements)
ISO/IEC/IEEE 42010:2022 (Architecture)
arc42 + C4 Model + Michael Nygard's ADR Format
Definition of Done Before completing work, verify ALL checkpoints:
✅ Legacy Migration (Phase 0 - if applicable):
✅ User Confirmation (Phase 1):
✅ Coordinator + Workers Invoked Sequentially (Phase 2):
✅ File Verification Complete:
✅ Global Cleanup Complete (Phase 4):
✅ Documentation Audit (Phase 5 - if selected):
✅ Summary Displayed (Phase 6):
✅ Error Handling (if applicable):
Output: Complete documentation system (CLAUDE.md + docs/ with README.md, documentation_standards.md, principles.md + presentation/ + optionally tests/) with global cleanup (no duplicates, no orphaned files, consolidated knowledge, validated cross-links)
Version: 8.1.0 (Added Documentation Standards section with NO_CODE, Stack Adaptation, Format Priority rules)
Last Updated: 2025-01-12
