ccprompts – claude code command collection

ccprompts is a comprehensive Claude Code plugin featuring 70+ commands across 12 development phases, 10 specialized AI agents, and a Dagger-based safety system. Install as a plugin for instant access to enterprise-grade workflows, or use the repository directly for customization. Commands include safety validation, the agent template system provides wizards for creating specialized subagents, and the multi-dimensional validation engine ensures quality and security. For more info on subagents, see subagents.

Key Features:

• 70+ production-ready slash commands organized by development phase
• 10 specialized AI agents for domain-specific expertise
• Dagger container isolation for safe command execution
• Multi-dimensional validation (structure, security, quality, performance)
• MCP integration and workflow automation capabilities
• Enterprise governance and multi-repo management tools

Development Notice: This repository is under active development. Commands may contain bugs, breaking changes can occur between versions, and the structure may evolve. Use with caution in production environments.

Quick start

Recommended Prerequisites

Install Dagger (not explicitly required, but recommended to enable containerized command execution):

Linux:

curl -fsSL https://dl.dagger.io/dagger/install.sh | BIN_DIR=$HOME/.local/bin sh

macOS:

brew install dagger/tap/dagger

Windows:

winget install Dagger.Cli

Installation

git clone https://github.com/ursisterbtw/ccprompts.git; and cd ccprompts; and bun i; and bun run validate

Adding a new command

1. Copy an existing command from .claude/commands/ to your desired phase folder.
2. Update the structure and content for your specific use case.
3. Include safety validation steps.
4. Run bun run validate before committing.

Creating specialized agents

The repository includes an agent creation system using templates/SUBAGENT_TEMPLATE.md:

1. Use the agent-template-wizard agent to create new specialized agents
2. Provide domain expertise and specific capabilities needed
3. Wizard fills template placeholders with consistent structure
4. Generated agents follow 7-step methodology with proper categorization
5. Agents include examples, color coding, and validation approaches

The template system ensures consistent agent structure while allowing domain-specific customization.

---

Installation as a Plugin

ccprompts can be installed as a Claude Code plugin, giving you instant access to all 70+ commands and 10 specialized agents in any project.

Quick Plugin Installation

# In Claude Code
/plugin marketplace add ursisterbtw/ccprompts
/plugin install ccprompts@ursisterbtw

Restart Claude Code and all commands will be available via /help.

Benefits of Plugin Installation

• All commands available in any project without manual copying
• Automatic updates when new versions are released
• Clean separation between your code and the command ecosystem
• Team-wide consistency through shared plugin configuration
• No need to clone the repository for each project

Team Installation

Add to your project's .claude/settings.json for automatic team-wide installation:

{
  "pluginMarketplaces": [
    "ursisterbtw/ccprompts"
  ],
  "plugins": [
    "ccprompts@ursisterbtw"
  ]
}

Local Development

For testing local changes or contributing:

# Run the test setup script
./scripts/test-plugin-local.sh

# Or manually
/plugin marketplace add /path/to/ccprompts
/plugin install ccprompts@ccprompts

For complete plugin documentation, installation methods, and troubleshooting, see PLUGIN.md.

---

Repository layout

.claude/
├── commands/          # ~70 commands across 12 phases (00-11)
│   ├── 00-initial-workflow/     # Project analysis and workflow (2 commands)
│   ├── 01-project-setup/        # Documentation, learning, MCP (3 commands)
│   ├── 02-development/          # Backup, debug, optimize, refactor (5 commands)
│   ├── 03-security/             # Security auditing and compliance (4 commands)
│   ├── 04-testing/              # Testing and troubleshooting (2 commands)
│   ├── 05-deployment/           # Deployment and CI/CD (4 commands)
│   ├── 06-collaboration/        # Code review and team workflow (4 commands)
│   ├── 07-utilities/            # Productivity and management tools (10 commands)
│   ├── 08-extras/               # Health checks and modernization (4 commands)
│   ├── 09-agentic-capabilities/ # MCP and agent orchestration (12 commands)
│   ├── 10-ai-native-development/ # AI-powered development tools (10 commands)
│   └── 11-enterprise-scale/     # Governance and multi-repo (8 commands)
├── agents/            # Agent templates and configurations
├── workflows/         # Automated workflow definitions
├── README.md
├── settings.json
└── command-registry.json

scripts/               # Safety system + validation utilities
templates/             # subagent template files
tests/                 # Jest testing suite
docs/                  # Multi-level documentation

---

Usage Examples

cat .claude/commands/09-agentic-capabilities/mcp-discover.md
cat .claude/commands/10-ai-native-development/ai-pair-program.md
cat .claude/commands/11-enterprise-scale/governance.md

/mcp-discover install filesystem --path=/project/data
/ai-pair-program advanced typescript --context-aware
/governance policy organization soc2 --enforce-automatically

Safe command execution

./scripts/safe-run.sh "command"

See SAFETY.md for full safety system documentation.

---

Community & Support

• Open an issue for bugs, feature requests, or questions
• Start a discussion by opening an issue for ideas and feedback
• See CONTRIBUTING.md for guidelines

---

Contributing

Pull requests are welcome. Please:

1. Run the validator and ensure no errors.
2. Follow conventional commit messages (feat: …, fix: …, etc.).
3. Keep commands clear and focused—avoid unnecessary complexity.

---

License

MIT. See LICENSE for details.

---