Install
$ agentstack add skill-severity1-claude-code-plugin-creator-plugin-creator ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Plugin Creator
Purpose
This skill guides the creation of Claude Code plugins - modular packages that extend Claude Code with custom slash commands, specialized agents, reusable skills, event hooks, and external service integrations. Use this skill to build plugins ranging from simple utilities to complex development toolkits.
When to Use This Skill
Apply this skill when the user wants to:
- Create a new Claude Code plugin from scratch
- Add components to an existing plugin (commands, agents, skills, hooks, MCP servers)
- Set up a plugin development environment
- Understand plugin structure and best practices
- Test or distribute a plugin
Common usage patterns:
- "Create a plugin with /deploy and /rollback commands"
- "Build a Python development toolkit plugin"
- "Add a code formatter hook to my existing plugin"
- "Create a plugin that integrates with the GitHub API"
- "Help me scaffold a new Claude Code plugin"
How to Use This Skill
Development Workflow
Follow the core development workflow:
- Understanding - Gather 2-3 concrete examples of how the plugin will be used
- Planning - Determine which components are needed (commands, agents, skills, hooks, MCP servers)
- Implementation - Execute
python scripts/init_plugin.pyto scaffold the structure - Customization - Customize generated components, remove unused ones, update metadata
- Testing - Install via local marketplace, validate, and iterate
For detailed procedures (Phases 0-5), example scenarios, and distribution guidance, consult references/development-workflow.md.
Component Types:
- Commands - User-invoked operations (
/deploy,/test) - Skills - Reusable workflows with bundled resources
- Agents - Automatic delegation to specialized tasks
- Hooks - Automatic responses to events
- MCP servers - External service integration
Consult references/component-specifications.md for detailed component guidance.
Iteration and Refinement
After initial implementation, improve the plugin through systematic iteration:
- Test activation - Verify commands appear in
/help, agents in/agents, skills activate on matching descriptions - Gather feedback - Have users test realistic scenarios and report issues
- Refine components - Update component files based on testing results
- Validate changes - Run
claude plugin validateafter each modification - Reinstall and retest - Use uninstall/install cycle to test changes:
``bash claude plugin uninstall claude plugin install @local-dev ``
- Monitor conflicts - Ensure commands/skills do not conflict with other plugins
- Document improvements - Track changes for version history
For detailed iteration procedures, consult references/development-workflow.md Phase 3.
Using Bundled Resources
This skill provides scripts, references, and templates to streamline plugin development:
scripts/init_plugin.py Execute this script directly via Bash to scaffold a new plugin. Do not load into context for typical usage.
python scripts/init_plugin.py [--path ]
Creates complete plugin structure with all component directories, plugin.json manifest, example files, local .dev-marketplace for testing, and README template. Requires Python 3.6+ (standard library only).
Load into context only when customization of initialization behavior is needed (e.g., modifying scaffold logic, changing default structure).
references/component-specifications.md
- Load for detailed specs on all component types
- Includes plugin manifest schema
- Reference when implementing commands, agents, skills, hooks, or MCP servers
references/plugin-json-examples.md
- Load for plugin.json structure reference (minimal, standard, advanced configurations)
- Quick lookup when scaffolding or updating manifest
- Grep patterns for targeted loading:
- "Minimal Plugin" | "Standard Plugin" | "Advanced Plugin" - configuration examples
references/component-examples.md
- Load for real-world component implementations
- Shows complete working examples for commands, agents, skills, hooks, and MCP servers
- Grep patterns for targeted loading:
- "/deploy" | "AWS Deployment" | "Testing Workflow" - specific component examples
- "Command:" | "Agent:" | "Skill:" | "Hook:" | "MCP:" - component type sections
references/pattern-library.md
- Load for advanced implementation patterns and edge cases
- Covers CRUD patterns, progressive loading, hook chains, tool restrictions, custom paths
- Grep patterns for targeted loading:
- "CRUD Plugin" | "Progressive Loading" | "Hook Chain" - implementation patterns
- "Monorepo" | "Non-Standard" | "Component-Specific Testing" - advanced topics
references/development-workflow.md
- Load for testing, validation, and distribution procedures
- Follow when setting up local development or preparing to publish
- Grep patterns for targeted loading:
- "Phase 0" | "Phase 1" | "Phase 2" | "Phase 3" | "Phase 4" | "Phase 5" - development phases
- "Single Command Plugin" | "Multi-Component" | "Event-Driven" - example scenarios
- "Best Practices" | "Useful Commands" - quick reference
references/troubleshooting.md
- Load when encountering validation errors or structural issues
- Contains common problems and actionable solutions
references/marketplace-guide.md
- Load only when ready to distribute the plugin
- Covers marketplace creation, git setup, and publishing
assets/ templates
- Copy as starting points when creating components
command-template.md- Slash command structureagent-template.md- Agent structureskill-template.md- Skill structurehooks-template.json- Hook configurationmcp-template.json- MCP server definitionsmarketplace-template.json- Marketplace structure
Templates are minimal, copy-ready files (not loaded into context).
Language and Style
Write all plugin content using:
- Imperative/infinitive form (verb-first instructions)
- Objective, instructional tone
- Third-person in YAML frontmatter descriptions
- Clear, actionable guidance
Avoid second-person language ("you should") - instead use "To accomplish X, do Y" format.
Additional Resources
For canonical, up-to-date documentation, see official Claude Code documentation.
Throughout the development process, ask clarifying questions to understand user needs. Focus on the most important questions first to avoid overwhelming the user.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: severity1
- Source: severity1/claude-code-plugin-creator
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.