skillhub-ai 0.4.0

The skill package manager for AI agents — install, compose, optimize, and bridge skills across Claude, Cursor, Codex, and Gemini

skillhub

The skill composer for AI agents.
Merge skills from any ecosystem into one expert skill — with AI-powered conflict resolution.

pip install skillhub-ai

---

# Three skills from three different companies → one expert
skillhub compose anthropic:claude-api openai:aspnet-core google:agent-platform-deploy -o cloud-expert

# Use it in Claude Code instantly
/cloud-expert

---

Why skillhub?

You install a python-patterns skill in Claude Code. You find a security-review skill in Cursor. You see a frontend-developer persona in agency-agents. They all live in different files, different formats, different tools.

The problem:

• No way to combine them — copy-pasting markdown is error-prone and tedious
• Duplicate sections silently contradict each other across files
• Skills from Anthropic, OpenAI, Microsoft, Google have no common interface
• Every agent (Claude Code, Cursor, Codex, Gemini) requires a different file format

skillhub solves this in one command:

skillhub compose anthropic:security-hardening openai:aspnet-core \
               agency-agents:frontend-developer python-patterns \
               -o my-expert --strategy ai

---

Features

🔀 Compose — The Only Tool That Merges Skills

Merge 2–10 skills from any source into one unified expert skill with automatic conflict detection. No other tool does this.

🤖 AI-Powered Conflict Resolution

When two skills define the same ## Error Handling section, --strategy ai sends both to Claude and gets back a single best-of-both version. No manual editing.

🌐 13 Ecosystems, One Interface

Pull from Anthropic, OpenAI, GitHub Copilot, Microsoft, Google, Vercel, and 8 more — using a simple ecosystem:skill-name prefix. All sources resolve to the same SKILL.md format.

🔍 Diff Before You Merge

See exactly which sections will conflict before composing. Green = unique to A, yellow = conflict, blue = unique to B.

📦 Pre-Built Expert Templates

5 production-ready combinations: fastapi-expert, fullstack-expert, ml-platform, pre-pr-reviewer, research-analyst. One command to expert.

🤝 Works With Every Agent

Claude Code (.claude/commands/), Cursor (.cursor/rules/), OpenAI Codex (AGENTS.md), Gemini CLI (.gemini/skills/). One skill, four agents.

---

Quick Start

# 1. Install
pip install skillhub-ai

# 2. Browse what's available
skillhub list
skillhub search "security"

# 3. Preview a skill before installing
skillhub info debug-agent

# 4. Install to Claude Code
skillhub install debug-agent

# 5. Install to ALL agents at once
skillhub install debug-agent --all-agents

# 6. Compose your first expert
skillhub compose --template fastapi-expert

---

Compose

[!IMPORTANT] No other tool does this. skillhub compose is the only way to merge multiple SKILL.md files into one unified expert skill with conflict detection and resolution.

# Simplest: registry skills
skillhub compose python-patterns security-review api-design -o fastapi-expert

# Cross-company: Anthropic + OpenAI + Google
skillhub compose anthropic:claude-api openai:aspnet-core google:agent-platform-deploy -o cloud-expert

# Mix everything: community + local + any ecosystem
skillhub compose agency-agents:frontend-developer ./my-standards.md skills.sh:react-expert -o our-expert

# Preview without writing
skillhub compose python-patterns security-review -o test --dry-run

What happens when you compose:

python-patterns  +  security-review  +  api-design
       │                  │                  │
       └──────────────────┴──────────────────┘
                          │
                    Fetch each source
                    Parse ## sections
                    Detect conflicts
                          │
              ┌───────────┴───────────┐
       first-wins (default)         ai (Claude)
       first skill wins         merges both intelligently
              └───────────┬───────────┘
                          │
            .claude/commands/fastapi-expert.md
                          │
                  /fastapi-expert ✓

---

AI-Powered Merge

pip install "skillhub-ai[ai]"
export ANTHROPIC_API_KEY=sk-...

skillhub compose python-patterns security-review -o secure-python --strategy ai

Strategy	Speed	On conflict
first-wins (default)	Instant	First skill's section kept, conflict logged
ai	~2s/conflict	Claude reads both, writes a unified best-of-both

---

Diff Before You Merge

skillhub diff python-patterns security-review

Comparing: python-patterns  vs  security-review
────────────────────────────────────────────────
✅ Only in python-patterns (6):   Type Annotations, Async Patterns, Dataclasses...
✅ Only in security-review  (5):  OWASP Top 10, Input Validation, Auth Patterns...
⚠️  Conflicts (2):               Error Handling, Code Review Checklist

---

Pre-Built Expert Templates

skillhub templates
skillhub compose --template fastapi-expert
skillhub compose --template ml-platform --strategy ai

Template	Skills	Best for
fastapi-expert	python-patterns + security-review + api-design	FastAPI backend
fullstack-expert	react-patterns + python-patterns + api-design	Full-stack
ml-platform	agent-builder + mle-workflow + llm-evaluator	ML engineers
pre-pr-reviewer	code-reviewer + security-review + test-writer	Code review
research-analyst	research-agent + rag-evaluator + doc-generator	Research

---

The 13 Ecosystems

[!TIP] Any prefix works in compose, diff, install, and info — mix freely.

skillhub compose \
  anthropic:claude-api \               # Anthropic official
  openai:aspnet-core \                 # OpenAI official
  copilot:acquire-codebase-knowledge \ # GitHub Copilot official
  microsoft:skill-creator \            # Microsoft official
  google:agent-platform-deploy \       # Google official
  addyosmani:code-review-and-quality \ # Addy Osmani
  scientific:astropy \                 # K-Dense AI
  antigravity:ab-test-setup \          # Antigravity/OpenClaw
  gamedev:godot-2d-movement \          # Game Dev Skills
  tech-leads:domain-analysis \         # Tech Leads Club
  agency-agents:frontend-developer \   # Agency Agents
  debug-agent \                        # skillhub registry
  -o the-ultimate-expert

Prefix	Ecosystem	Repo
(no prefix)	skillhub registry	chandrudp29/skillhub
anthropic:name	Anthropic official	anthropics/skills
openai:name	OpenAI / Codex official	openai/skills
copilot:name	GitHub Copilot official	github/awesome-copilot
microsoft:name	Microsoft official	microsoft/skills
google:name	Google official	google/skills
skills.sh:name	Vercel skills.sh	vercel-labs/skills
agency-agents:name	Agency Agents	msitarzewski/agency-agents
addyosmani:name	Addy Osmani	addyosmani/agent-skills
scientific:name	K-Dense AI	K-Dense-AI/scientific-agent-skills
antigravity:name	Antigravity	sickn33/antigravity-awesome-skills
gamedev:name	Game Dev Skills	gamedev-skills/awesome-gamedev-agent-skills
tech-leads:name	Tech Leads Club	tech-leads-club/agent-skills
claude:name	Your Claude Code project	.claude/commands/
cursor:name	Your Cursor project	.cursor/rules/
codex:name	Your Codex project	AGENTS.md
gemini:name	Your Gemini CLI project	.gemini/skills/
github:owner/repo/path	Any public GitHub file	raw.githubusercontent.com
./path/to/file.md	Local file	—

---

Install to Any Agent

skillhub install debug-agent                  # Claude Code (default)
skillhub install debug-agent --agent cursor   # Cursor
skillhub install debug-agent --agent codex    # OpenAI Codex
skillhub install debug-agent --agent gemini   # Gemini CLI
skillhub install debug-agent --all-agents     # all 4 at once
skillhub install debug-agent --dry-run        # preview only

Agent	Where it's written	How to invoke
Claude Code	.claude/commands/<name>.md	/<name> in chat
Cursor	.cursor/rules/<name>.mdc	Always active
OpenAI Codex	AGENTS.md (with markers)	Auto-loaded
Gemini CLI	.gemini/skills/<name>.md	Always active

---

Cross-Ecosystem Recipes

# Ultimate Python expert — 3 sources, AI-merged
skillhub compose python-patterns security-review api-design -o python-expert --strategy ai

# Enterprise frontend — community + production-grade + patterns
skillhub compose agency-agents:frontend-developer addyosmani:code-review-and-quality \
  react-patterns -o enterprise-frontend --strategy ai

# Cloud AI expert — Anthropic + Google + Microsoft official skills
skillhub compose anthropic:claude-api google:agent-platform-deploy \
  microsoft:skill-creator -o cloud-ai-expert

# Scientific Python — K-Dense AI + skillhub registry
skillhub compose scientific:astropy python-patterns rag-evaluator -o research-scientist

# Pre-PR gate (template + AI merge)
skillhub compose --template pre-pr-reviewer --strategy ai

---

Create & Publish Your Own Skill

skillhub init our-coding-standards    # scaffold SKILL.md template
# edit SKILL.md with your team's conventions
skillhub install our-coding-standards/  # test locally
skillhub publish our-coding-standards/  # open PR to registry

SKILL.md format

---
name: our-coding-standards
description: Backend coding standards for our team
version: 1.0.0
tags: [python, standards, backend]
agents: [claude, cursor, codex, gemini]
---

## When to Use
Apply when writing, reviewing, or refactoring any backend code.

## Core Rules
- All functions must have type annotations
- Use dataclasses over plain dicts for structured data

## Examples
...

The SKILL.md format is Anthropic's open standard — supported by 32+ AI coding tools.

---

All Commands

Expand full command reference

# Discovery
skillhub list                              # browse all 26 skills
skillhub list --installed                  # skills in this project
skillhub search <query>                    # search by keyword
skillhub search <query> --tag <tag>        # filter by tag
skillhub info <name>                       # full preview

# Install & Manage
skillhub install <name>                    # Claude Code (default)
skillhub install <name> --agent cursor     # Cursor
skillhub install <name> --all-agents       # all 4 agents
skillhub install <name> --dry-run          # preview without writing
skillhub install <name> --overwrite        # replace existing
skillhub uninstall <name>                  # remove
skillhub update                            # update all installed

# Compose
skillhub compose <a> <b> ... -o <name>             # merge (first-wins)
skillhub compose <a> <b> ... -o <name> --strategy ai  # Claude AI merge
skillhub compose --template <name>                 # pre-built template
skillhub compose --template <name> --strategy ai   # template + AI merge
skillhub compose <a> <b> -o <name> --dry-run       # preview only
skillhub templates                                 # list 5 templates

# Cross-ecosystem (any prefix works)
skillhub compose anthropic:security openai:aspnet-core -o expert
skillhub diff anthropic:claude-api openai:aspnet-core

# Diff
skillhub diff <a> <b>                      # compare section by section

# Create & Publish
skillhub init <name>                       # scaffold new SKILL.md
skillhub publish <path>                    # PR to registry

Command	What it does
skillhub diff <a> <b>	Compare two skills section by section
skillhub compose <a> <b> -o <name>	Merge (first-wins on conflicts)
skillhub compose ... --strategy ai	Merge with Claude AI resolution
skillhub compose --template <name>	Use a pre-built expert template
skillhub templates	List all built-in templates
skillhub search <query>	Search by name, description, or tag
skillhub list	Browse all 26 skills
skillhub list --installed	Skills installed in this project
skillhub info <name>	Full details and preview
skillhub install <name>	Install for Claude Code
skillhub install <name> --all-agents	Install for all 4 agents
skillhub install <name> --dry-run	Preview without writing
skillhub uninstall <name>	Remove a skill
skillhub update	Update all installed skills
skillhub init <name>	Scaffold a new SKILL.md
skillhub publish <path>	Submit to registry via PR

---

Available Skills (34)

Browse all 26 skills

Research & Analysis

Skill	Description
research-agent	Multi-source deep research with synthesis and citations
rag-evaluator	Evaluate RAG pipelines: faithfulness, relevance, hallucination
llm-evaluator	Build evaluation frameworks for LLM outputs

Code Quality

Skill	Description
code-reviewer	Five-axis review: correctness, security, performance, readability, maintainability
debug-agent	Systematic root cause analysis — feedback loop first
refactor-agent	Safe, incremental refactoring with verification
test-writer	Write meaningful tests that actually catch bugs
security-review	OWASP-based security auditing

Documentation

Skill	Description
doc-generator	README, API docs, docstrings, changelogs
pr-summarizer	Generate PR descriptions: what, why, how, risks

Development Patterns

Skill	Description
python-patterns	Modern Python best practices (typing, async, dataclasses)
react-patterns	React 18+ patterns (hooks, suspense, server components)
fastapi-patterns	Production FastAPI patterns
api-design	REST API conventions and best practices
sql-agent	Write, optimize, and explain SQL with query plans

DevOps

Skill	Description
docker-agent	Dockerfile optimization, multi-stage builds
git-workflow	Branching strategies, commit conventions

AI / ML Engineering

Skill	Description
agent-builder	Build LangGraph agents: tools, memory, streaming, multi-agent
prompt-optimizer	Diagnose and improve underperforming prompts
mle-workflow	Production ML workflow: data, training, deployment
cost-tracker	Track and optimize LLM API costs

Career

Skill	Description
yc-job-tracker	Daily AI/ML job tracking at YC startups
career-ops	Global AI/ML job hunt — USA, Europe, Singapore, Dubai

Web & Frontend

Skill	Description
typescript-patterns	Strict types, utility types, generics, discriminated unions, branded types
nextjs-patterns	App Router, Server Components, data fetching, caching, streaming

Architecture & Infrastructure

Skill	Description
system-design	RESDAC framework, scalability patterns, database selection, CAP trade-offs
cicd-agent	GitHub Actions, secrets, blue-green/canary/rolling deploys, release automation
kubernetes-agent	Manifests, resource sizing, HPA, health probes, networking, troubleshooting
data-pipeline	Idempotency, incremental loads, Airflow DAGs, data quality checks

Performance & APIs

Skill	Description
performance-optimizer	Profiling methodology, Python bottlenecks, N+1 fixes, caching strategy
openai-patterns	Model selection, structured outputs, function calling, retry logic, cost control

Community (via agency-agents)

Skill	Description
frontend-developer	Senior frontend developer expertise
backend-architect	Senior backend architecture and system design
penetration-tester	Security testing methodology

Skills from all 13 external ecosystems are also available via their prefix.

---

Contributing

skillhub is a community registry — every skill you add helps thousands of developers.

1. Fork this repo
2. Add your skill: skillhub init <name> then edit skills/<name>/SKILL.md
3. Add an entry to registry/index.json
4. Open a PR

Or in one command: skillhub publish <name>/

What makes a good skill? Focused on one domain, has ## When to Use and ## Core Rules, works well when composed with other skills, gives concrete examples over vague advice.

---

Author

Built by Chandrashekar DP — Senior AI/ML Engineer, Bengaluru.

If skillhub saves you time, a ⭐ helps others find it.

---

License

MIT — use it, fork it, build on it.

---

skillhub · pip install skillhub-ai · PyPI · Issues · Discussions