Agent skill
Documentation Updater
Automatically updates CLAUDE.md and related documentation when new features, configurations, or best practices are discovered during development sessions
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/data/documentation-updater-lucasdoreac-mutuapix-workspace
SKILL.md
Documentation Updater Skill
Overview
This skill implements a self-improving documentation system for the MutuaPIX project. It automatically:
- Detects when new information is discovered during sessions
- Updates CLAUDE.md with new commands, patterns, and configurations
- Maintains skills documentation (SKILL.md files)
- Tracks version history of documentation changes
Core Principle: Progressive Disclosure
Documentation follows the progressive disclosure pattern:
- Quick Reference - Essential commands and patterns in CLAUDE.md
- Detailed Guides - In-depth explanation in SKILL.md files
- Context Files - Supporting documentation (README, audit reports, etc.)
Auto-Update Triggers
The system should update documentation when:
1. New Commands Discovered
Trigger: Running a bash command that solves a problem and should be remembered
Example:
# During session, you run:
ssh root@138.199.162.115 'cd /var/www/mutuapix-frontend-production && rm -rf .next && npm run build'
# Auto-update CLAUDE.md:
## Quick Commands > Frontend > Build
# Clear cache and rebuild
ssh root@138.199.162.115 'cd /var/www/mutuapix-frontend-production && rm -rf .next && npm run build'
2. New Configuration Patterns
Trigger: Discovering environment variable requirements or configuration settings
Example:
# Discovered: NEXT_PUBLIC_NODE_ENV required for production
# Auto-update: CLAUDE.md > Environment Variables section
3. Security Vulnerabilities Found
Trigger: Identifying security issues during code review or testing
Example:
# Found: authStore initializes with mock user
# Auto-create: SECURITY_AUDIT_[DATE].md
# Auto-update: CLAUDE.md > Security > Known Issues
4. New Skills Created
Trigger: Creating a new SKILL.md file
Example:
# Created: .claude/skills/pix-validation/SKILL.md
# Auto-update: CLAUDE.md > Available Skills section
5. Workflow Improvements
Trigger: Discovering better way to perform existing task
Example:
# Old way: Manual file transfer
# New way: rsync with specific flags
# Auto-update: CLAUDE.md > Deployment section
Update Workflow
Step 1: Detect Update Trigger
interface UpdateTrigger {
type: 'command' | 'config' | 'security' | 'skill' | 'workflow';
source: string; // Where was this discovered?
content: string; // What should be documented?
priority: 'low' | 'medium' | 'high' | 'critical';
relatedFiles: string[]; // Which files are affected?
}
Step 2: Determine Update Location
Decision Tree:
Is it a quick command/reference?
YES → Update CLAUDE.md > Quick Commands
NO → Continue
Is it detailed technical knowledge?
YES → Update or create SKILL.md
NO → Continue
Is it a security issue?
YES → Create SECURITY_AUDIT_[DATE].md + update CLAUDE.md
NO → Continue
Is it project-specific documentation?
YES → Update README.md or docs/
NO → Log for future review
Step 3: Format Update
CLAUDE.md Updates:
## Section Title
### Subsection (if needed)
**Description:** Brief explanation of what/why
**Command/Pattern:**
```bash
# Comment explaining what this does
command here
Related Files: path/to/file.ext
Added: 2025-10-16 (Track when added)
**SKILL.md Updates:**
```markdown
## Version History
- **1.1.0** (2025-10-16): [Description of update]
- Added: Feature X
- Fixed: Issue Y
- Updated: Section Z
Step 4: Verify Update
Checklist:
- Is the information accurate?
- Is it concise (CLAUDE.md sections should be <200 words)?
- Is it actionable (commands can be copy-pasted)?
- Is it versioned (date added, version number)?
- Are related files cross-referenced?
Step 5: Commit Changes
# Auto-commit documentation updates
git add CLAUDE.md .claude/skills/
git commit -m "docs: auto-update from session [SESSION_ID]
- Added: [brief description]
- Updated: [sections modified]
- Source: [what triggered update]
Co-Authored-By: Claude <noreply@anthropic.com>"
CLAUDE.md Structure
Required Sections
# CLAUDE.md
**Last Updated:** YYYY-MM-DD
**Location:** `/Users/lucascardoso/Desktop/MUTUA/`
**Auto-Update Enabled:** ✅
---
## 🎯 Project Overview
[Brief description, tech stack]
## 🚀 Quick Start Commands
[Most common commands for dev/deploy]
## 🏗️ Architecture
[High-level system design]
## 🔐 Security
[Critical security considerations]
## 📋 Available Skills
[List of SKILL.md files with descriptions]
## ⚙️ Configuration
[Environment variables, important settings]
## 🐛 Troubleshooting
[Common issues and solutions]
## 📚 Documentation References
[Links to detailed docs]
## 🔄 Version History
[Track major CLAUDE.md changes]
Keep It Concise
❌ BAD (Too verbose):
The MutuaPIX platform uses a comprehensive authentication system
that leverages Laravel Sanctum on the backend for secure token-based
authentication while the frontend utilizes Next.js with Zustand for
state management and localStorage for token persistence...
✅ GOOD (Concise):
## Authentication
**Stack:** Laravel Sanctum (backend) + Zustand (frontend)
**Token Lifetime:** 24 hours
**Login Flow:** CSRF token → POST /api/v1/login → Store JWT
**Quick Test:**
```bash
curl https://api.mutuapix.com/api/v1/health
Details: See .claude/skills/authentication-management/SKILL.md
## Skills Discovery
When Claude needs to find a skill:
```bash
# List all available skills
ls -la .claude/skills/
# Search skills by keyword
grep -r "PIX validation" .claude/skills/*/SKILL.md
# Read specific skill
cat .claude/skills/pix-validation/SKILL.md
Auto-discovery: Skills are automatically discovered from:
.claude/skills/(project-level)~/.claude/skills/(personal)- Plugin-provided skills
Update Examples
Example 1: New Environment Variable Discovered
Trigger:
# During debugging, discovered NEXT_PUBLIC_NODE_ENV is required
Auto-Update CLAUDE.md:
## Environment Variables
### Frontend Production (.env.production on VPS)
```bash
NEXT_PUBLIC_NODE_ENV=production # ⚠️ CRITICAL: Required for security
NEXT_PUBLIC_API_URL=https://api.mutuapix.com
NEXT_PUBLIC_USE_AUTH_MOCK=false
Why NEXT_PUBLIC_NODE_ENV is Critical:
process.env.NODE_ENVis undefined in Next.js client-side code- Mock authentication relies on this for environment detection
- Missing this variable = security bypass in production!
Added: 2025-10-16
Related Files: frontend/src/lib/env.ts, frontend/src/stores/authStore.ts
### Example 2: New Deployment Command
**Trigger:**
```bash
# Discovered that clearing .next cache is required for env var changes
ssh root@138.199.162.115 'cd /var/www/mutuapix-frontend-production && rm -rf .next && npm run build && pm2 restart mutuapix-frontend'
Auto-Update CLAUDE.md:
## Deployment
### Frontend Deployment (Environment Variable Changes)
**⚠️ IMPORTANT:** When updating `.env.production`, must clear cache before rebuild!
```bash
# 1. Update .env.production on VPS
ssh root@138.199.162.115 'echo "NEXT_PUBLIC_NODE_ENV=production" >> /var/www/mutuapix-frontend-production/.env.production'
# 2. Clear cache + rebuild + restart
ssh root@138.199.162.115 'cd /var/www/mutuapix-frontend-production && rm -rf .next && npm run build && pm2 restart mutuapix-frontend'
# 3. Verify
curl -I https://matrix.mutuapix.com/login
Why Clear Cache:
Next.js caches compiled bundles. Without clearing .next/, old environment variables persist.
Added: 2025-10-16
### Example 3: Security Issue Found
**Trigger:**
```typescript
// Discovered: authStore has default mock user
// File: frontend/src/stores/authStore.ts:91-96
user: devLocalUser, // ❌ VULNERABILITY
Auto-Create: SECURITY_AUDIT_2025_10_16.md (full report)
Auto-Update CLAUDE.md:
## Security
### Known Issues
**🔴 CRITICAL: Default Mock User in authStore** (Found: 2025-10-16)
**Issue:** `authStore` initializes with authenticated mock admin user by default.
**Risk:** If localStorage is empty, user appears logged in without credentials.
**Fix:** See `AUTHENTICATION_AUDIT_REPORT.md` Section "Remediation Plan > Phase 1"
**Status:** 🟡 Pending fix
**Related Files:**
- `frontend/src/stores/authStore.ts:91-96`
- `.claude/skills/authentication-management/SKILL.md`
Self-Improvement Loop
Phase 1: Discovery
- Claude encounters new information during session
- Flags it as potential documentation update
Phase 2: Validation
- Is this information accurate?
- Is it useful for future sessions?
- Where should it be documented?
Phase 3: Update
- Edit CLAUDE.md or create/update SKILL.md
- Follow formatting guidelines
- Add version history entry
Phase 4: Verification
- Read updated documentation
- Verify it's clear and actionable
- Check for consistency with existing docs
Phase 5: Commit
- Create git commit with descriptive message
- Tag with "docs:" prefix for auto-tracking
Metrics & Tracking
Documentation Health Metrics
interface DocHealthMetrics {
claudeMdSize: number; // Bytes (target: <50KB)
skillsCount: number; // Number of SKILL.md files
lastUpdated: Date; // Most recent update
staleSections: string[]; // Sections >90 days old
brokenLinks: string[]; // Dead file references
todoItems: number; // Unresolved TODOs in docs
}
Update Frequency
Target:
- CLAUDE.md: Updated after every significant session (new feature, bug fix, deployment)
- SKILL.md: Updated when skill scope changes (version bump)
- Security docs: Updated immediately when vulnerability found
Quality Checks
Before updating CLAUDE.md, verify:
- Total file size <50KB (concise documentation)
- No duplicate information across sections
- All commands tested and work
- All file paths are valid
- No sensitive data (passwords, API keys)
- Version history entry added
Integration with Claude Sessions
At Session Start
- Read CLAUDE.md to understand project context
- Load relevant SKILL.md files based on user's task
- Check for stale documentation (>30 days old)
During Session
- Note when new information is discovered
- Flag for documentation update
- Continue with task (don't interrupt flow)
At Session End
- Review flagged documentation updates
- Apply updates to CLAUDE.md and/or SKILL.md
- Commit changes with descriptive message
- Summary of documentation changes for user
User Prompt for Updates
When user says:
- "Update CLAUDE.md with this" → Apply update immediately
- "Remember this for next time" → Add to appropriate doc
- "This is important" → Flag as high priority for CLAUDE.md
- "Document this workflow" → Create or update SKILL.md
Best Practices
DO:
✅ Keep CLAUDE.md concise (<50KB total) ✅ Use bullet points over paragraphs ✅ Include copy-pasteable commands ✅ Cross-reference detailed docs in SKILL.md ✅ Add dates to new entries ✅ Version bump SKILL.md when updating
DON'T:
❌ Duplicate information across files ❌ Include verbose explanations in CLAUDE.md ❌ Commit broken/untested commands ❌ Leave TODOs unresolved for >7 days ❌ Remove historical information (archive instead)
Related Files
Core Documentation:
CLAUDE.md- Main project guide (you are here).claude/skills/*/SKILL.md- Detailed skill documentationREADME.md- Public-facing project readme
Audit Reports:
AUTHENTICATION_AUDIT_REPORT.md- Security audit (2025-10-16)VPS_AUDIT_REPORT.md- Infrastructure auditCLEANUP_EXECUTION_REPORT.md- VPS cleanup log
Legacy:
WORKFLOW_RULES_FOR_CLAUDE.md- Git workflow rulesdocs/- Additional project documentation
Version History
- 1.0.0 (2025-10-16): Initial documentation updater skill
- Defined auto-update triggers
- Established update workflow
- Created documentation structure guidelines
- Implemented self-improvement loop
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
Didn't find tool you were looking for?