Shared Components

shared/skills/documentation/
└── SKILL.md

---
name: documentation
description: Technical documentation writer
---
 
# Documentation Skill
 
Write clear, comprehensive technical documentation.
 
## Guidelines
 
- Use clear, concise language
- Include code examples
- Document edge cases
- Keep docs close to code

shared/skills/testing/
├── SKILL.md
└── config.json

---
name: testing
description: Test writing assistant
---
 
# Testing Skill
 
Help write comprehensive tests.
 
## Supported Frameworks
 
See config.json for framework-specific patterns.

{
  "frameworks": {
    "pytest": {
      "pattern": "test_*.py",
      "assertions": "assert"
    },
    "jest": {
      "pattern": "*.test.ts",
      "assertions": "expect"
    }
  }
}

shared/agents/debugger.md

---
name: debugger
description: Autonomous debugging agent
---
 
# Debugger Agent
 
I am a debugging agent. When activated:
 
1. Analyze the error or unexpected behavior
2. Form hypotheses about root causes
3. Add strategic logging/debugging
4. Test hypotheses systematically
5. Propose and verify fixes
 
## Approach
 
- Start with reproduction
- Isolate the problem
- Binary search for root cause
- Validate fix doesn't introduce regressions

shared/agents/refactorer.md

---
name: refactorer
description: Code refactoring specialist
---
 
# Refactoring Agent
 
I specialize in improving code quality without changing behavior.
 
## Capabilities
 
- Extract methods/functions
- Simplify conditionals
- Remove duplication
- Improve naming
- Optimize imports
 
## Safety Rules
 
- Always maintain existing behavior
- Run tests after each change
- Make small, incremental changes
- Document significant changes

shared/hooks/lint-check/
├── hook.json
└── run.sh

{
  "name": "lint-check",
  "event": "pre-commit",
  "command": "./run.sh"
}

#!/bin/bash
echo "Running linter..."
# Add your lint command here

shared/hooks/notify/
├── hook.json
└── notify.py

{
  "name": "notify",
  "event": "post-task",
  "command": "python notify.py"
}

shared/rules/security.md

---
name: security
description: Security guidelines
---
 
# Security Rules
 
## Never
 
- Hardcode credentials or secrets
- Use eval() or exec() with user input
- Disable SSL verification
- Log sensitive data
 
## Always
 
- Validate and sanitize inputs
- Use parameterized queries
- Implement proper authentication
- Follow least privilege principle

shared/rules/code-style.md

---
name: code-style
description: Code style guidelines
---
 
# Code Style Rules
 
## Naming
 
- Use descriptive names
- camelCase for variables/functions
- PascalCase for classes
- UPPER_SNAKE for constants
 
## Structure
 
- Keep functions under 50 lines
- Maximum 3 levels of nesting
- One class per file (generally)

// shared/agents/security-audit/agent.json
{
  "name": "security-audit",
  "description": "Comprehensive security audit agent",
  "dependencies": {
    "skills": ["vulnerability-scan", "code-review"],
    "rules": ["security"]
  }
}

// shared/skills/full-stack-review/skill.json
{
  "name": "full-stack-review",
  "description": "Full stack code review",
  "dependencies": {
    "skills": ["frontend-review", "backend-review", "database-review"]
  }
}

cldpm add agent:security-audit --to my-project

✓ Added agents/security-audit to my-project
  ✓ skills/vulnerability-scan (dependency)
  ✓ skills/code-review (dependency)
  ✓ rules/security (dependency)

cldpm add skill:documentation --to my-project
cldpm add agent:debugger --to my-project
cldpm add rule:security --to my-project

cldpm add agent:security-audit --to my-project --no-deps

for project in web-app api-server worker; do
  cldpm add skill:testing --to $project
  cldpm add rule:security --to $project
done

cldpm remove skill:testing --from my-project

cldpm get my-project --format json | jq '.shared'

# Edit shared skill
vim shared/skills/testing/SKILL.md
 
# Changes are immediately available in all linked projects
# No need to run cldpm sync