15 KiB
title, description, hide
| title | description | hide | ||
|---|---|---|---|---|
| Agent Skills & Plugins for Claude Code, Codex, Gemini CLI & 10 More AI Tools | 345 production-ready agent skills, 78 installable plugins, and 90+ slash commands across 17 domains — engineering, product, marketing, compliance, finance, and research. Works with Claude Code, OpenAI Codex, Gemini CLI, Cursor, Hermes Agent, Mistral Vibe, OpenClaw, and 6 more AI coding tools. Open source, MIT licensed, zero dependencies. |
|
Open source · MIT · 17 domains · 13 AI tools
Agent Skills
Give your AI coding agent real domain expertise. Every skill is a self-contained package of workflows, checklists, Python tools, and reference knowledge that your agent follows autonomously — install one command, ship better work. { .hero-subtitle }
Get Started{ .md-button .md-button--primary } Browse Skills{ .md-button } GitHub :fontawesome-brands-github:{ .md-button }
Works with
What Is an Agent Skill?
An agent skill is a portable package of expertise your AI assistant can load on demand. Instead of re-explaining your standards in every prompt, a skill gives the agent a structured workflow it follows the same way every time — whether that's reviewing a pull request, designing a pricing model, or preparing an ISO 27001 audit.
Every skill in this library follows the same simple anatomy:
skill-name/
├── SKILL.md # The playbook — workflows, rules, decision frameworks
├── scripts/ # Python CLI tools (stdlib-only, no pip installs)
├── references/ # Curated domain knowledge the agent can consult
└── assets/ # Ready-to-use templates for your team
No API keys, no external services, no dependencies between skills. Copy a folder — or install a plugin — and it works.
- Install Add the marketplace to Claude Code, or run one sync script for Codex, Gemini CLI, Cursor, and 9 more tools.
- Invoke Call a slash command like
/cs:deal-review, or just mention the skill in a prompt — the agent loads the playbook. - Ship The agent works through the skill's checklists and tools, producing consistent, reviewable output every time.
What's Inside
-
:material-toolbox:{ .lg .middle } 345 Skills
Production-ready playbooks across 17 domains — from code review and RAG architecture to pricing strategy, clinical study design, and markdown-to-HTML publishing. Each ships with workflows, Python tools, and reference docs.
-
:material-puzzle-outline:{ .lg .middle } 78 Plugins
One-command installable bundles for Claude Code — install a whole domain or a single skill. Sync scripts cover Codex CLI, Gemini CLI, Hermes Agent, Mistral Vibe, and OpenClaw.
-
:material-robot:{ .lg .middle } 90+ Agents
Multi-skill orchestrators with distinct personas — engineering leads, C-suite advisors, research routers, and compliance auditors that combine skills for complex work.
-
:material-console:{ .lg .middle } 90+ Slash Commands
Instant workflows you run by name — sprint planning, PRDs, OKRs, deal reviews, SLO design, chaos experiments, and market research, straight from your terminal.
-
:material-language-python:{ .lg .middle } 570+ Python Tools
Deterministic CLI scripts bundled with skills — all standard library, zero pip installs, no LLM calls. Scoring, validation, and analysis that runs anywhere Python runs.
-
:material-account-group:{ .lg .middle } 3 Personas
Role-based identities — Startup CTO, Growth Marketer, Solo Founder — with curated skill loadouts, judgment frameworks, and distinct communication styles.
-
:material-sitemap:{ .lg .middle } Orchestration
A lightweight protocol for coordinating personas, skills, and agents on work that crosses domain boundaries — launches, audits, and strategic sprints.
-
:material-swap-horizontal:{ .lg .middle } 13-Tool Support
Write once, run everywhere. A single conversion script adapts every skill to the native format of Cursor, Aider, Windsurf, Kilo Code, OpenCode, Augment, and Antigravity.
-
:material-chat-outline:{ .lg .middle } 6 Custom GPTs
Use Agent Skills directly in ChatGPT with zero setup — Solo Founder, SEO Audit, Content Strategy, CTO Advisor, and more.
Skills by Domain
Seventeen domains cover the full lifecycle of building a product and running a company — engineering, go-to-market, operations, compliance, and research.
-
:material-cog:{ .lg .middle } Engineering — Core
Architecture, frontend, backend, fullstack, QA, DevOps, SecOps, AI/ML, data engineering, Playwright testing, self-improving agent
-
:material-lightning-bolt:{ .lg .middle } Engineering — Advanced
Agent designer, RAG architect, MCP server builder, CI/CD pipelines, SLO architect, chaos engineering, security auditing, tech debt tracking
-
:material-bullseye-arrow:{ .lg .middle } Product
Product manager toolkit, agile PO, UX research, discovery, analytics, experiment design, SaaS scaffolding, Apple HIG
-
:material-bullhorn:{ .lg .middle } Marketing
Content, SEO, AEO, CRO, paid channels, growth, launch strategy — 8 specialist pods with bundled Python analytics tools
-
:material-star-circle:{ .lg .middle } C-Level Advisory
Full C-suite advisors (CEO to General Counsel), founder-mode boardroom, decision logging, culture and strategy frameworks
-
:material-shield-check:{ .lg .middle } Regulatory & Quality
ISO 13485, MDR 2017/745, FDA, ISO 27001, GDPR, CAPA, risk management, quality documentation
-
:material-shield-lock:{ .lg .middle } Compliance OS
Audit-prep orchestrator for ISO 13485, ISO 27001, SOC 2, GDPR, FDA QSR, EU AI Act, and ISO 42001 readiness
-
:material-clipboard-check:{ .lg .middle } Project Management
Senior PM, scrum master, Jira and Confluence experts, Atlassian admin with bundled remote MCP
-
:material-trending-up:{ .lg .middle } Business & Growth
Customer success, sales engineering, revenue operations, contracts and proposals
-
:material-cog-outline:{ .lg .middle } Business Operations
Process mapping, vendor management, capacity planning, internal comms, knowledge ops, procurement
-
:material-handshake:{ .lg .middle } Commercial
Pricing strategy, deal desk, partnerships, channel economics, commercial policy, RFP response, forecasting
-
:material-currency-usd:{ .lg .middle } Finance
Financial analysis, DCF valuation, budgeting, forecasting, SaaS metrics (ARR, MRR, churn, LTV)
-
:material-magnify:{ .lg .middle } Research
Literature review, grants, patents, entity dossiers, syllabi, NotebookLM automation — with a hybrid orchestrator
-
:material-flask:{ .lg .middle } Research Operations
Clinical study design, R&D program finance, market sizing and surveys, product research methodology
-
:material-lightning-bolt-outline:{ .lg .middle } Productivity
Brain-dump capture, inbox setup and triage, reflection journal, session handoff, market-first decision making
-
:material-language-html5:{ .lg .middle } Markdown to HTML
Turn markdown into beautiful single-file HTML — long-form documents, code reviews, and slide decks with your brand
-
:material-web:{ .lg .middle } Landing Pages
Single-file HTML landing page generator with four design styles and a brand palette validator
Why Teams Use This Library
-
:material-clock-fast:{ .lg .middle } Zero dependencies
Every Python tool uses the standard library only. No pip installs, no API keys, no configuration. Works anywhere Python runs.
-
:material-shield-lock:{ .lg .middle } Security first
A built-in security auditor scans any skill for malicious code, data exfiltration, and prompt injection before you install it.
-
:material-rocket-launch:{ .lg .middle } One-command install
A plugin marketplace for Claude Code, sync scripts for Codex, Gemini, Hermes, and Vibe, and a converter for seven more tools.
-
:material-puzzle:{ .lg .middle } Self-contained
Each skill is independent — no cross-dependencies, no conflicts. Install one or all; they work in isolation.
-
:material-devices:{ .lg .middle } Multi-platform
Native support for 13 AI coding tools. Write once, convert to any tool's format automatically.
-
:material-check-decagram:{ .lg .middle } Production-grade
Structured workflows with validation checkpoints — not generic advice. Each skill covers an end-to-end process with named deliverables.
Quick Install
=== "Claude Code"
```bash
# Add the marketplace
/plugin marketplace add alirezarezvani/claude-skills
# Install any skill bundle
/plugin install engineering-skills@claude-code-skills
```
=== "OpenAI Codex"
```bash
npx agent-skills-cli add alirezarezvani/claude-skills --agent codex
```
=== "Gemini CLI"
```bash
git clone https://github.com/alirezarezvani/claude-skills.git
cd claude-skills && ./scripts/gemini-install.sh
```
=== "Hermes Agent"
```bash
git clone https://github.com/alirezarezvani/claude-skills.git
cd claude-skills
python scripts/sync-hermes-skills.py --verbose
# Skills appear in /skills and /<skill-name> automatically
```
=== "Mistral Vibe"
```bash
git clone https://github.com/alirezarezvani/claude-skills.git
cd claude-skills
./scripts/vibe-install.sh
# Skills install to ~/.vibe/skills/claude-skills/; same SKILL.md standard
```
=== "Cursor / Windsurf / Aider"
```bash
git clone https://github.com/alirezarezvani/claude-skills.git
cd claude-skills
./scripts/convert.sh --tool cursor # or windsurf, aider
./scripts/install.sh --tool cursor --target /path/to/project
```
Full Install Guide{ .md-button .md-button--primary } Multi-Tool Setup{ .md-button }
Guides
Tool-specific walkthroughs for getting the most out of the library:
- Best Claude Code Plugins & Skills — the 20 plugins to start with, by use case
- Agent Skills for OpenAI Codex CLI — install and invoke skills in Codex
- Gemini CLI Skills & Plugins Guide — setup, indexing, and usage for Gemini CLI
- Cursor Agent Skills & Rules Guide — convert skills to Cursor's rules format
- OpenClaw Skills Guide — one-line install for OpenClaw workspaces