agentleFS
bitovibitovi/ai-enablement-prompts/.claude/skills/create-plugin/SKILL.md

create-plugin

Skill121 starsChanged 6 months agoMITView on GitHub

Use this skill when the user asks to "create a plugin", "add a plugin", "make a new plugin", "build a plugin", or wants to package skills into an installable plugin for this marketplace. Guides creating a properly structured plugin under plugins/.

14 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions14 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 14

    1. ToolJet/ToolJet★ 41ksame name
    2. rundeck/rundeck★ 6.4ksame name
    3. mistralai/mistral-vibe★ 5.3ksame name
    4. datavane/tis★ 1.4ksame name
    5. marcus/sidecar★ 1.1ksame name
    6. alexsmedile/hormozi-skills★ 208same name
    7. lbedner/aegis-stack★ 146same name
    8. npow/oh-my-claude★ 4same name
    9. rzx007/openharness-ts★ 3same name
    10. rzx007/Vykor★ 3same name

    4 more not listed

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

The file

Show its rules and text

We split this skill into the rules it gives an agent

14 rules. 0 are common, 6 are distinctive, and 8 are very specific to this project.

  • 0

    Common

    Most of the 12 edited versions we read keep these.

  • 6

    Distinctive

    This author's own choices. Worth a look.

  • 8

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Ask the user (or infer from context):

    line 37
  2. Distinctive:

    No slashes, colons, or namespace prefixes — these cause silent load failures

    line 63
  3. Very specific:

    If you set version, you must bump it on every release or users won't receive updates

    line 69
  4. Very specific:

    Do NOT set version in both plugin.json and marketplace.json — plugin.json wins silently

    line 70
  5. Very specific:

    Every SKILL.md must have YAML frontmatter:

    line 86
  6. Distinctive:

    Use numbered steps for sequential workflows

    line 101
  7. Distinctive:

    Add a checklist at the end for verification

    line 103
  8. Very specific:

    Keep SKILL.md under 500 lines — move detailed reference material to separate files

    line 104
  9. Very specific:

    Create plugins/<plugin-name>/README.md with this template:

    line 127
  10. Very specific:

    Add the plugin to .claude-plugin/marketplace.json in the plugins array:

    line 178
  11. Distinctive:

    Structure check: Verify all required files exist:

    line 194
  12. Very specific:

    JSON validity: Ensure plugin.json and marketplace.json are valid JSON:

    line 201
  13. Very specific:

    Frontmatter check: Every SKILL.md must start with --- and have name and description fields

    line 207
  14. Distinctive:

    Check this table before creating a new plugin — the skill may belong in an existing one.

    line 231

“Common” is measured: kept by at least 6 in 10 of the 12 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Create Plugin

Help the user create a new plugin in this repo's marketplace.

What Is a Plugin?

A plugin is an installable package of AI agent skills distributed through this marketplace. Users install plugins to get specialized capabilities for their tech stack. Each plugin lives in plugins/<plugin-name>/ and is registered in the marketplace catalog.

Repo Structure

ai-enablement-prompts/
├── .claude-plugin/
│   └── marketplace.json          # Marketplace catalog — lists all plugins
├── .github/
│   └── plugin/
│       └── marketplace.json      # Symlink to .claude-plugin/marketplace.json
└── plugins/
    └── <plugin-name>/
        ├── .claude-plugin/
        │   └── plugin.json       # Plugin identity (name, description, author)
        ├── skills/
        │   └── <skill-name>/
        │       └── SKILL.md      # Skill instructions + YAML frontmatter
        ├── instructions/          # (optional) Scoped instruction files
        │   └── *.instructions.md
        └── README.md             # Install & usage docs

Step 1: Determine Plugin Scope

Ask the user (or infer from context):

  • What tech stack does this plugin target?
  • What skills should it include?
  • Does it require any MCP servers (Figma, Playwright, Atlassian, etc.)?
  • Is it a standalone plugin or does it complement other plugins in the marketplace?

Grouping principle: Plugins are organized by tech stack and MCP dependencies, not by workflow stage. Users install only what applies to their stack.

Step 2: Create plugin.json

Create plugins/<plugin-name>/.claude-plugin/plugin.json:

{
  "name": "<plugin-name>",
  "description": "One-line description of what this plugin provides.",
  "author": {
    "name": "Bitovi",
    "url": "https://www.bitovi.com"
  }
}

Naming Rules

  • name must be plain kebab-case: lowercase letters, numbers, hyphens only (max 64 chars)
  • No slashes, colons, or namespace prefixes — these cause silent load failures
  • Examples: code, react, playwright, figma-react, trpc-prisma

Version Pinning

  • Omit version to use the git commit SHA as version (every commit = automatic update)
  • If you set version, you must bump it on every release or users won't receive updates
  • Do NOT set version in both plugin.json and marketplace.json — plugin.json wins silently

Step 3: Create Skills

Each skill gets its own folder under plugins/<plugin-name>/skills/<skill-name>/:

plugins/<plugin-name>/skills/<skill-name>/
├── SKILL.md          # Main instructions (required)
├── reference/        # Background knowledge, specs, checklists (optional)
├── steps/            # Step-by-step sub-instructions (optional)
└── examples/         # Example inputs/outputs (optional)

SKILL.md Frontmatter

Every SKILL.md must have YAML frontmatter:

---
name: skill-name
description: Use this skill when the user asks to "...". Include specific trigger phrases and situations.
---
  • name: kebab-case, matches the folder name
  • description: How the agent decides when to auto-load this skill — include trigger phrases, keywords, and situations

Writing Effective Skills

  • State the goal clearly at the top
  • Use numbered steps for sequential workflows
  • Include "When to Use" and "When NOT to Use" sections
  • Add a checklist at the end for verification
  • Keep SKILL.md under 500 lines — move detailed reference material to separate files
  • Reference supporting files with relative markdown links: [checklist.md](reference/checklist.md)

Step 4: Create Instructions (Optional)

For scoped coding conventions that apply to specific file patterns, create instruction files:

plugins/<plugin-name>/instructions/<name>.instructions.md

With frontmatter specifying the scope:

---
applyTo: '**/packages/server/**'
---

Instructions are automatically applied when the agent works on files matching the applyTo glob pattern. Use these for coding standards, not for task workflows (those should be skills).

Step 5: Create README.md

Create plugins/<plugin-name>/README.md with this template:

# <Plugin Name>

<One paragraph description>

## Skills

| Skill | Description |
|-------|-------------|
| `skill-name` | What it does |

## Prerequisites

- List any required MCP servers
- List any required tech stack (React, Zod, tRPC, etc.)

## Install

### Claude Code

\```bash
claude plugin marketplace add bitovi/ai-enablement-prompts --scope project
claude plugin install <plugin-name>@bitovi-ai-enablement --scope project
\```

Commit `.claude/settings.json` so teammates get the plugin automatically.

### VS Code Copilot

Add to `.vscode/settings.json` (commit this file):

\```json
{
  "chat.plugins.marketplaces": ["bitovi/ai-enablement-prompts"]
}
\```

Then install from: **Extensions view** → search `@agentPlugins` → find `<plugin-name>` → Install

### GitHub Copilot CLI

\```bash
copilot plugin marketplace add bitovi/ai-enablement-prompts
copilot plugin install <plugin-name>@bitovi-ai-enablement
\```

Step 6: Register in Marketplace

Add the plugin to .claude-plugin/marketplace.json in the plugins array:

{
  "name": "<plugin-name>",
  "source": "./plugins/<plugin-name>",
  "description": "Same one-line description as plugin.json"
}

.github/plugin/marketplace.json is a symlink to .claude-plugin/marketplace.json — do not edit it separately. Claude Code and the GitHub Copilot CLI both resolve to the same file.

Step 7: Verify

Run these checks:

  1. Structure check: Verify all required files exist:
    ls plugins/<plugin-name>/.claude-plugin/plugin.json
    ls plugins/<plugin-name>/README.md
    ls plugins/<plugin-name>/skills/*/SKILL.md
  2. JSON validity: Ensure plugin.json and marketplace.json are valid JSON:
    cat plugins/<plugin-name>/.claude-plugin/plugin.json | python3 -m json.tool
    cat .claude-plugin/marketplace.json | python3 -m json.tool
  3. Frontmatter check: Every SKILL.md must start with --- and have name and description fields
  4. Local test (Claude Code):
    claude --plugin-dir ./plugins/<plugin-name>
  5. Validate (Claude Code):
    claude plugin validate ./plugins/<plugin-name>

Existing Plugins in This Marketplace

PluginDescriptionRequired MCPs
creating-promptsMeta-skills for creating new agent skills (Claude + Copilot)—
codeSpec workflow, signatures, instruction-generation, document-feature, create-skill—
reactModlet pattern, component extraction, registry, reuse, validation—
react-mockZod sample data, data model management, feature implementation—
figma-reactFigma-to-React lifecycle: design, implement, sync, Code ConnectFigma
playwrightE2E testing, debugging, responsive design, visual diff, pixel-perfectPlaywright
trpc-prismaCross-package types, AppRouter inference, package instructions—

Check this table before creating a new plugin — the skill may belong in an existing one.

Common Mistakes

  • ❌ Forgetting to add the plugin to marketplace.json
  • ❌ Using uppercase, slashes, or colons in name (causes silent failures)
  • ❌ Setting version in plugin.json and then forgetting to bump it
  • ❌ Putting task workflows in instruction files (use skills instead)
  • ❌ Putting coding conventions in skills (use instruction files instead)
  • ❌ Creating a new plugin when the skill fits in an existing one

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.