Skip to content
Back to skills

Create Plugin

ASecurity

Scaffolds a new top-level agent plugin directory. NOT for scaffolding single skills (use `create-skill`) and NOT for adding MCP integrations to existing plugins (use `create-mcp-integration`).

  • 7 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added September 2, 2026
ai-agentspythonbash

Works with

  • mcp

Security analysis

A100/100

Pro scans all 20 files and shows the line behind each finding

Scanned October 3, 2026

npx -y skills add richfrem/agent-plugins-skills --skill create-plugin --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Create Plugin?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Create Plugin
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/richfrem-create-plugin-agent-plugins-skills/badge)](https://www.skillsdirectory.com/skills/richfrem-create-plugin-agent-plugins-skills)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: create-plugin
plugin: agent-scaffolders
description: >
  Scaffolds a new top-level agent plugin directory. NOT for scaffolding single skills (use `create-skill`) and NOT for adding MCP integrations to existing plugins (use `create-mcp-integration`).
argument-hint: "[plugin-name]"
allowed-tools: Bash, Read, Write
---

# Create Plugin (`create-plugin`)

Scaffolds a complete agent plugin with multi-platform manifests and standard directory layout.

## Contents

- [Constraints](#constraints)
- [Quick start](#quick-start)
- [Workflow](#workflow)
- [Verification](#verification)
- [References](#references)

## Constraints

- **Clean metadata**: `.claude-plugin/plugin.json` must never declare `skills`, `agents`, or `hooks` arrays (auto-discovered).
- **Author object**: `author` must be an object `{"name": "...", "email": "..."}`, never a string.
- **Hub-first scripts**: Reusable scripts live in `plugins/<plugin>/scripts/` and symlink into skills.
- **File-level symlinks only**: Directory symlinks violate repository policy.
- **No cross-plugin imports**: Never import Python code from another plugin.

## Quick start

```bash
# Scaffold plugin root directory structure
mkdir -p plugins/<plugin-name>/{.claude-plugin,skills,references,scripts,tests}
```

## Workflow

1. **Scaffold Root**: Create directory structure with `.claude-plugin/plugin.json` and `README.md`.
2. **Configure Multi-Platform Manifests**:
   - `.claude-plugin/plugin.json`: Minimal metadata format with author object.
   - `plugin.yaml`: Hermes compatibility manifest (declaring kind, platforms, tools).
3. **Scaffold Components**: Use `create-skill` for skills and `create-hook` for hooks.
4. **Symlink Hygiene**: Link hub scripts and references to skill folders.
5. **Audit**: Run `audit-plugin` to verify structure, schemas, and symlinks.

## Verification

```bash
# Validate plugin structure compliance
python3 scripts/audit_plugin_structure.py plugins/<plugin-name>
# Validate hook schemas if hooks present
python3 scripts/validate_hook_schema.py plugins/<plugin-name>/hooks/hooks.json
```

## References

- [create-plugin-guide.md](references/create-plugin-guide.md) — Detailed workflows and manifest templates.
- [acceptance-criteria.md](references/acceptance-criteria.md) — Plugin creation acceptance criteria.
- [fallback-tree.md](references/fallback-tree.md) — Troubleshooting and error recovery tree.

Files in this skill

  • SKILL.md7 KB
  • acceptance-criteria.md533 B
  • evals/evals.json892 B
  • evals/results.tsv417 B
  • fallback-tree.md1.4 KB
  • references/ADRs/001_cross_plugin_script_dependencies.md62 B
  • references/ADRs/002_multi_skill_script_centralization.md63 B
  • references/ADRs/003_plugin_skill_resource_sharing_via_mirrored_folder_structure_and_file_level_symlinks.md113 B
  • references/ADRs/004_self_contained_plugins_no_cross_plugin_dependencies.md81 B
  • references/ADRs/005_plugin_separation_of_concerns_and_loose_coupling.md78 B
  • references/ADRs/006_python_native_plugin_bootstrap_installer.md70 B
  • references/acceptance-criteria.md42 B
  • references/fallback-tree.md36 B
  • references/hitl-interaction-design.md46 B
  • references/pattern-decision-matrix.md46 B
  • references/patterns/action-forcing-output-with-deadline-attribution.md82 B
  • references/patterns/adversarial-objectivity-constraint.md69 B
  • references/patterns/anti-pattern-vaccination.md59 B
  • references/patterns/anti-symptom-triage.md54 B
  • references/patterns/artifact-embedded-execution-audit-trail.md74 B

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.