document-sync
Keeps project documentation in sync with current code and architecture.
Install
mkdir -p .claude/skills/document-sync && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/3126" && unzip -o skill.zip -d .claude/skills/document-sync && rm skill.zipInstalls to .claude/skills/document-sync
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
A robust skill that analyzes your app's actual codebase, tech stack, configuration, and architecture to ensure ALL documentation is current and accurate. It never assumes—always verifies and compares the live system with every documentation file to detect code-doc drift and generate actionable updates.Key capabilities
- →Detects documentation-code drift using system introspection
- →Identifies undocumented configuration variables
- →Generates PRs for documentation updates
- →Assigns trust scores based on verifiable code evidence
How it works
It parses source code and configuration files to extract structural truth, then compares this data against documentation files to find inconsistencies.
Inputs & outputs
When to use document-sync
- →Check docs for accuracy against code
- →Identify undocumented configuration variables
- →Generate documentation updates for new API routes
About this skill
Document Sync Skill for Claude Code
Overview
This skill provides comprehensive documentation synchronization by analyzing your actual codebase, tech stack, configuration, and architecture to ensure all documentation remains current and accurate. It operates on a "verify, never assume" principle, performing end-to-end comparisons against real code and dependencies to detect documentation drift automatically.
Core Principles
- Verify, Never Assume: Documentation is only updated after end-to-end comparison against real, current code and dependencies
- Deep System Introspection: Checks languages, frameworks, APIs, DBs, config, code patterns, and deployment targets
- Doc Trust Score: Every doc gets a trust rating based on evidence found in the codebase and config files
Quick Start
Natural Language Commands:
- "Run document sync analysis on my project"
- "Update docs so every API route matches the code"
- "Check if any doc mentions features no longer in the app"
- "List undocumented config variables"
Slash Commands:
/docs-sync— Runs analysis and outputs a complete report on outdated/inaccurate sections/docs-update-strict— Suggests only changes verified by found code/config evidence/docs-rewrite-missing— Fills gaps for undocumented features based on live source/docs-patch-pr— Makes all doc changes as a PR, zero surprises
Operation Workflow
Phase 1: Full System Scan
Execute the system scanning script to analyze your codebase:
# Run comprehensive system scan
python3 .claude/skills/document-sync/scripts/system_scan.py
This phase automatically:
- Tech Stack Detection - Parse
package.json,requirements.txt,pom.xml, etc. for frameworks, DBs, and build tools - Project Structure Analysis - Map
src/,lib/,app/, andconfig/directories for real usage - Architecture Detection - Identify data flows, auth methods, sync strategies, and cloud/on-prem deployment
Key Outputs: system_state.json and code_features_report.md
Phase 2: Documentation Verification
Execute the documentation verification script:
# Verify all documentation against code
python3 .claude/skills/document-sync/scripts/verify_docs.py
This phase:
- Docs Content Inventory - List all Markdown, docs/, and README files
- Technical Validation - Cross-reference each doc claim with real code/config
- Relevance & Quality Scoring - Rate every doc: current, needs update, obsolete, or missing
Key Outputs: doc_verification_report.md and docs_update_plan.md
Phase 3: Automated or Assisted Document Updating
Based on the verification results, choose your update approach:
# Dry run - see what would change
python3 .claude/skills/document-sync/scripts/update_docs.py --mode=dry-run
# Suggest mode - get suggestions for review
python3 .claude/skills/document-sync/scripts/update_docs.py --mode=suggest
# Auto-update with backups
python3 .claude/skills/document-sync/scripts/update_docs.py --mode=auto-update
Configuration
Create .claude/document-sync-skill.yml in your project root:
mode: suggest # suggest | dry-run | auto-update
protected_docs:
- README.md
- docs/security.md
review_required_for:
- section removal
- breaking doc changes
output_reports:
- doc_verification_report.md
- docs_update_plan.md
Safety & Controls
- Never deletes or rewrites without firm code evidence
- Always creates a diff and backup before change
- User sets conservatism level: dry-run, suggest-only, or auto-update
- All changes are traceable—committed separately with actionable summaries
Output Examples
doc_verification_report.md
| Doc Section | Status | Evidence Found | Action |
|---|---|---|---|
/docs/api.md GET /users | Current | Endpoint in code (src/routes/) | Keep |
/docs/db-setup.md PouchDB | Outdated | No mention in codebase | Remove/Update |
/README.md Feature X | Missing | Found in src/feature-x.ts | Add details |
docs_update_plan.md
- Remove obsolete DB doc:
/docs/db-setup.md(no longer in code) - Add section to
/README.mdfor unlisted Feature X - Rewrite
/docs/config.mdsection on env vars (now using.env.local) - Confirm all code-blocks in
/docs/usage.mdstill run
Resources
This skill includes specialized scripts and references for comprehensive document synchronization:
scripts/
- system_scan.py - Comprehensive codebase analysis and tech stack detection
- verify_docs.py - Documentation verification and validation against code
- update_docs.py - Automated document updating with safety controls
references/
- tech_detection_patterns.md - Patterns for identifying frameworks and technologies
- doc_validation_rules.md - Rules for validating documentation accuracy
- update_templates.md - Templates for common documentation updates
assets/
- report_templates/ - Markdown templates for verification and update reports
- config_schemas/ - JSON schemas for configuration validation
Safety Checklist
- Full project scan completed
- All doc claims matched to real code/config
- Backups created before overwriting docs
- No deletions/rewrites without clear code evidence
- All updates and removals require user review
- All code examples in docs validated with real source
Success Metrics
- 100% doc claims match code/config truth
- 0 obsolete/incorrect references in doc set
- All new features in code are documented within 2 days
- Doc update effort reduced by >60% for every release
MANDATORY USER VERIFICATION REQUIREMENT
Policy: No Fix Claims Without User Confirmation
CRITICAL: Before claiming ANY issue, bug, or problem is "fixed", "resolved", "working", or "complete", the following verification protocol is MANDATORY:
Step 1: Technical Verification
- Run all relevant tests (build, type-check, unit tests)
- Verify no console errors
- Take screenshots/evidence of the fix
Step 2: User Verification Request
REQUIRED: Use the AskUserQuestion tool to explicitly ask the user to verify the fix:
"I've implemented [description of fix]. Before I mark this as complete, please verify:
1. [Specific thing to check #1]
2. [Specific thing to check #2]
3. Does this fix the issue you were experiencing?
Please confirm the fix works as expected, or let me know what's still not working."
Step 3: Wait for User Confirmation
- DO NOT proceed with claims of success until user responds
- DO NOT mark tasks as "completed" without user confirmation
- DO NOT use phrases like "fixed", "resolved", "working" without user verification
Step 4: Handle User Feedback
- If user confirms: Document the fix and mark as complete
- If user reports issues: Continue debugging, repeat verification cycle
Prohibited Actions (Without User Verification)
- Claiming a bug is "fixed"
- Stating functionality is "working"
- Marking issues as "resolved"
- Declaring features as "complete"
- Any success claims about fixes
Required Evidence Before User Verification Request
- Technical tests passing
- Visual confirmation via Playwright/screenshots
- Specific test scenarios executed
- Clear description of what was changed
Remember: The user is the final authority on whether something is fixed. No exceptions.
When not to use it
- →When working on projects without any existing documentation
- →For codebases utilizing highly obfuscated or generated languages
Prerequisites
Limitations
- →Cannot detect functional logic changes that don't manifest in code structure
- →May generate false positives for legacy docs that intentionally omit code details
How it compares
It uses mechanical verification against live code rather than relying on natural language analysis which can overlook technical implementation changes.
Compared to similar skills
document-sync side by side with the closest alternatives in the catalog.
| Skill | Installs | Updated | Safety | Difficulty |
|---|---|---|---|---|
| document-sync (this skill) | 2 | 7mo | Review | Intermediate |
| documentation-review | 11 | 4mo | No flags | Beginner |
| docs-review | 10 | 7mo | No flags | Beginner |
| workthrough | 10 | 8mo | Review | Beginner |
Try saying
Example prompts that trigger this skill in your AI assistant.
More by ananddtyagi
View all by ananddtyagi →You might also like
documentation-review
stacklok
Reviews documentation for factual accuracy
docs-review
metabase
Review documentation changes for compliance with the Metabase writing style guide. Use when reviewing pull requests, files, or diffs containing documentation markdown files.
workthrough
bear2u
Automatically document all development work and code modifications in a structured workthrough format. Use this skill after completing any development task, bug fix, feature implementation, or code refactoring to create comprehensive documentation.
claude-md-improver
anthropics
Audit and improve CLAUDE.md files in repositories. Use when user asks to check, audit, update, improve, or fix CLAUDE.md files. Scans for all CLAUDE.md files, evaluates quality against templates, outputs quality report, then makes targeted updates. Also use when the user mentions "CLAUDE.md maintenance" or "project memory optimization".
anti-slop
rand
Comprehensive toolkit for detecting and eliminating "AI slop" - generic, low-quality AI-generated patterns in natural language, code, and design. Use when reviewing or improving content quality, preventing generic AI patterns, cleaning up existing content, or enforcing quality standards in writing, code, or design work.
agent-md-refactor
davila7
Refactor bloated AGENTS.md, CLAUDE.md, or similar agent instruction files to follow progressive disclosure principles. Splits monolithic files into organized, linked documentation.