home / skills / cdeistopened / opened-vault / manage-skills

manage-skills skill

/.claude/skills/manage-skills

This skill helps you create, update, and maintain Claude Code skills following established design principles, ensuring consistent, self-contained modules.

npx playbooks add skill cdeistopened/opened-vault --skill manage-skills

Review the files below or copy the command above to add this skill to your agents.

Files (1)
SKILL.md
1.5 KB
---
name: manage-skills
description: Manages Claude Code skills - creates, updates, and maintains skills following established design principles. Use when the user asks to create a skill, update a skill, refactor a skill, or wants to teach Claude a new capability.
version: 3
---

# Skill Management

## Skill Design Principles

### 1. SKILL.md is Self-Contained
- Contains ALL information needed to use the skill
- Should be as minimal as possible while conveying complete information
- No need for separate README, USAGE, INSTALL, or CHANGELOG files

### 2. Single Script Design
- Optimize for ONE script per skill (not multiple scripts)
- Use command-line parameters for different operations
- Pattern: `./script.sh <command> [arguments]`

### 3. Minimal File Structure
```
skill-name/
├── SKILL.md          # Required - complete documentation
└── script.sh         # Optional - single CLI if needed
```

## SKILL.md Structure

Required frontmatter:
```yaml
---
name: skill-name
description: What it does and when to use it. Use when [triggers].
version: 1
---
```

## Creating a New Skill

1. Create directory in `.claude/skills/skill-name/`
2. Create SKILL.md with frontmatter and documentation
3. Optionally add a shell script for automation
4. Make scripts executable with `chmod +x`

## Rendering Custom UI

For rendering interactive HTML interfaces in chat, use the **create-interface** skill which provides comprehensive documentation on the `mcp__noetect-ui__render_ui` tool.

Overview

This skill manages Claude Code skills by creating, updating, and maintaining skill packages following established design principles. It streamlines authoring so each skill is self-contained, simple to run, and easy to maintain. Use it to add new capabilities, refactor existing skills, or teach Claude a new behavior.

How this skill works

The skill enforces a single, self-contained documentation file per skill that contains all usage and configuration details. It prefers one executable script per skill that accepts command-style arguments for operations. The tool automates directory creation, documentation templating, permission settings, and lightweight validation checks.

When to use it

  • Creating a new skill to teach Claude a capability
  • Updating or refactoring an existing skill for consistency
  • Converting multi-file skills into a single-script form
  • Automating repetitive setup tasks and permission fixes
  • Preparing skills for sharing or deployment in a workspace

Best practices

  • Keep the documentation file minimal but complete—include name, short description, triggers, and version
  • Use one executable script per skill and implement command-style arguments (e.g., ./script.sh <command> [args])
  • Place each skill in its own directory with only the documentation and the script
  • Make scripts executable (chmod +x) and include basic input validation
  • Favor small, focused skills over large multipurpose ones for easier maintenance

Example use cases

  • Create a new skill directory with templated documentation and a sample CLI script
  • Refactor a skill that currently uses multiple files into a single-script layout with consolidated docs
  • Update the skill manifest and bump version after adding a new command
  • Automate the chmod step for all scripts in a workspace
  • Validate that a skill’s documentation includes required fields before publishing

FAQ

What should the single documentation file include?

Include the skill name, a concise description with triggers, usage instructions, commands, and a version number.

Why prefer one script per skill?

One script simplifies invocation, distribution, and maintenance; command arguments handle multiple operations without extra files.