agent-orchestration-rules
Rules for background agent execution - no polling, no TaskOutput, trust system reminders
SKILL.md
| Name | agent-orchestration-rules |
| Description | Rules for background agent execution - no polling, no TaskOutput, trust system reminders |
Claude Code Plugin Marketplace
A curated collection of Claude Code plugins, agents, skills, and tools for enhanced productivity. This marketplace provides professional-grade components for coding, documentation, and data analysis workflows.
Quick Start
Install the Marketplace
/plugin marketplace add Ven0m0/claude-config
Install Individual Plugins
# Core plugins
/plugin install coding-assistant@claude-config-marketplace
/plugin install conserve@claude-config-marketplace
/plugin install prompt-improver@claude-config-marketplace
/plugin install config-wizard@claude-config-marketplace
/plugin install dependency-blocker@claude-config-marketplace
/plugin install skill-authoring@claude-config-marketplace
# Security and tooling
/plugin install claude-praetorian@claude-config-marketplace
/plugin install moderntools@claude-config-marketplace
/plugin install dynamic-mcp-router@claude-config-marketplace
Use the Claude Config Pack Directly
Copy the claude/ directory to your Claude Code configuration:
cp -r claude/ ~/.claude/
This gives you 12 agents and 32 skills immediately.
Available Plugins (13 Total)
coding-assistant
Advanced coding assistant with code review, debugging, and refactoring capabilities.
/code-review- Comprehensive code reviews with security and performance checks/debug- Systematic debugging and root cause analysis/refactor- Code refactoring for better structure and maintainability- Automatic code formatting after edits (Biome, ruff, gofmt, rustfmt)
conserve
Resource optimization and performance monitoring toolkit for efficient Claude Code workflows.
/bloat-scan- Identify dead code, duplication, and documentation bloat/optimize-context- Optimize context usage and token efficiency/analyze-growth- Analyze codebase growth patterns/ai-hygiene-audit- Audit AI-generated code quality/unbloat- Remove unnecessary code and improve efficiency- Maximum Effective Context Window (MECW) principle
prompt-improver
Intelligent prompt optimization. Enriches vague prompts with research-based clarifying questions.
- Automatic prompt clarity evaluation
- Research-based clarifying questions (1-6 questions)
- Zero overhead for clear prompts
- 31% token reduction through skill-based architecture
config-wizard
Interactive wizard to create new Claude Code plugins and skills.
/config-wizard:cmd-init- Initialize a new slash command/config-wizard:cmd-review- Review an existing slash command
dependency-blocker
Blocks Claude from accessing dependency directories to reduce token usage.
- Blocks
node_modules,.git,dist,build,vendor,target,.venv,venv - Blocks Bash, Read, Glob, and Grep operations targeting excluded directories
skill-authoring
Framework and templates for creating new Claude Code skills.
claude-praetorian
Security guardrails and policy enforcement for Claude Code sessions.
moderntools
Enforces modern tool substitutions: fd over find, rg over grep, bun over npm, etc.
dynamic-mcp-router
Dynamic MCP server routing for flexible tool integration.
plugin-validator
Validates plugin structure and compliance against marketplace standards.
skills-eval
Skill evaluation and benchmarking framework.
skills_performance-optimization
Performance tuning for Claude Code skills.
claude-code-lsps
LSP integrations for Claude Code (go-to-definition, find-references, etc.).
Claude Config Pack
The claude/ directory is a standalone config pack with:
12 Sub-Agents (claude/agents/):
| Agent | Purpose |
|---|---|
ci-cd-expert | CI/CD pipeline design and troubleshooting |
code-explorer | Codebase analysis and architecture mapping |
code-simplifier | Refactoring for clarity and maintainability |
docker-specialist | Dockerfile optimization and security hardening |
general-purpose | Default multi-step task agent |
language-expert | Multi-language: Bash, Python, JS, TS, Rust |
maintenance | Cleanup, tech debt, DX improvements |
merge-supervisor | Git merge conflict resolution |
optimizer | Context and LLM optimization |
prd | Product Requirements Documents |
reverse-engineer | Binary analysis (authorized use only) |
skill-auditor | SKILL.md compliance audits |
32 Skills (claude/skills/):
ast-grep-search, bash-optimizer, code-antipatterns-analysis, code-execution,
data-formats, git-cli-agentic, github, hooks-configuration, javascript,
json-repair, linter-autofix, llm-boost, lsp-enable, mcp-builder, moai,
modern-tool-substitution, prd, python-project-development, ralph-planner,
render-output, repomix, ruff, rust, self-reflection, sequential-thinking,
strategic-compact, svg, toon-formatter, typescript, using-tmux-for-interactive-commands,
uv, vulture-dead-code
Repository Structure
claude-config/
├── AGENTS.md # Primary AI config (symlinked as CLAUDE.md, GEMINI.md)
├── README.md # This file
├── SETUP.md # Installation guide
├── CHANGELOG.md # Version history
│
├── claude/ # Claude Code config pack
│ ├── agents/ # 12 sub-agent definitions
│ ├── skills/ # 32 reusable skills
│ ├── hooks/ # Auto-running hooks (format, lint, MCP)
│ ├── docs/ # Reference documentation
│ └── settings.json # Permissions and model config
│
├── plugins/ # 13 installable plugins
├── gemini/ # Gemini CLI config
├── cursor/ # Cursor editor rules
├── copilot-cli/ # GitHub Copilot CLI config
├── opencode/ # OpenCode references
└── .github/
├── copilot-instructions.md # Copilot guardrails
└── workflows/ # CI/CD automation
Requirements
- Claude Code CLI (latest version)
- Git
rg(ripgrep) - required by hooksgh- GitHub CLI (for PR/issue workflows)- Optional:
fd,eza,ast-grep,jq,bun,uv
Documentation
- AGENTS.md - Complete agent, skill, and configuration reference
- SETUP.md - Detailed installation and configuration
- CHANGELOG.md - Version history
- claude/docs/ - Prompt patterns, hooks, LSP integration, TOON format
Contributing
Contributions welcome:
- Add new plugins or skills
- Improve existing agents
- Report issues or suggest features: GitHub Issues
License
MIT - see LICENSE