930+ Free AI Skills for Your Coding Assistant
We just open-sourced the largest AI skills library: 930+ production-grade skills and 31 agent personas that work across Claude Code, OpenAI Codex, Gemini CLI, GitHub Copilot, and Perplexity Computer.
One library. Five platforms. Zero setup. Get it on GitHub.
The Problem: Your AI Assistant Is Underperformingβ
You're using Claude Code or Codex to write code, review PRs, and debug issues. It works. But you know it could do more.
The gap isn't the model β it's the instructions. Without structured skill instructions, your AI assistant:
- Gives generic advice instead of domain-specific guidance
- Doesn't know your stack's best practices (Terraform, K8s, FastAPI, React patterns)
- Can't play specialized roles (security auditor, database architect, SRE)
- Produces output that needs heavy editing instead of being ready to use
Scattered prompt collections don't solve this. They're untested, single-platform, and organized by vibes instead of function.
What We Builtβ
The FAOS Skills Marketplace is a structured, categorized, cross-platform skills library:
| What | Count |
|---|---|
| Unique skills | 930+ |
| Agent personas | 31 |
| Platforms | 5 (Claude Code, Codex, Gemini CLI, Copilot, Perplexity) |
| Categories | 12+ (Engineering, Cloud, Security, Data/AI, Product, Growth...) |
| Industry coverage | 21 verticals |
| License | Apache 2.0 |
Every skill has structured metadata β name, description, tags, and platform-specific formatting. These aren't prompts. They're skill instructions that give your AI assistant reliable, repeatable expertise.
Install in 10 Secondsβ
Claude Code:
cp skills/cowork/code-reviewer/SKILL.md .claude/skills/
OpenAI Codex:
cp -r skills/codex/ .agents/skills/
Gemini CLI:
cp skills/gemini/commands/faos-skill-*.toml .gemini/commands/
That's it. No API keys. No configuration. No dependencies.
What's Insideβ
Engineering & DevOps (280+ skills)β
Code review, architecture patterns, CI/CD automation, Kubernetes manifests, Terraform modules, Docker optimization, database migrations, API design, testing workflows β the skills a senior engineering team uses daily.
Cloud & Infrastructure (120+ skills)β
Azure AI/ML (40+ skills covering every Azure AI SDK), AWS patterns, GCP services, Databricks, Snowflake, monitoring with Prometheus/Grafana/Datadog β deep cloud coverage, not surface-level wrappers.
Security & Compliance (60+ skills)β
OWASP Top 10, NIST CSF, SOC2 controls, HIPAA, PCI-DSS v4, threat modeling, SAST configuration, incident response playbooks, vulnerability scanning, secrets management β built by security engineers, not compliance checkboxes.
Data & AI/ML (100+ skills)β
LangChain, LangGraph, RAG implementation, vector database engineering (Qdrant, Pinecone, FAISS, Chroma), fine-tuning (LoRA, QLoRA, GRPO), MLOps, prompt engineering patterns, model evaluation β the full AI/ML stack.
31 Agent Personas β Not "Act As" Promptsβ
Each agent persona is a full role definition with:
- Communication style β how the agent talks
- Decision patterns β how it evaluates trade-offs
- KPIs β what it measures success by
- Vocabulary β domain-specific terminology
- Principles β what it believes
The CEO agent doesn't just "act as a CEO." It has a strategic framework, evaluates decisions against business impact, and communicates with executive clarity. The Security Engineer agent thinks in threat models and compliance frameworks. The Data Scientist agent reasons about statistical validity and experiment design.
11 C-Suite agents. 6 Engineering. 5 Data/AI. 3 Product. 3 GTM. 2 Delivery. 1 Quality.
Why We Open-Sourced Thisβ
We're the team behind FAOS β an enterprise agentic operating system that powers real business operations across 22 industries. These skills are extracted from production use.
We open-sourced the skills library because:
- Skills should be portable. You shouldn't be locked into one AI platform to get good instructions.
- Quality matters. Production-tested skills are better than crowd-sourced prompt collections.
- The real value is deeper. Skills give your AI instructions. The FAOS Platform gives your AI understanding β dynamic ontology context, memory, governance, and multi-agent orchestration. The skills library shows you the quality. The platform shows you the depth.
Skills vs. Platform β What's the Difference?β
| Skills Marketplace (Free) | FAOS Platform | |
|---|---|---|
| 930+ skills | Yes | Yes |
| 31 agent personas | Yes | Yes |
| 5 platform formats | Yes | Yes |
| Static instructions | Yes | Dynamic, context-aware |
| 7-phase execution engine | β | Yes |
| 3-layer memory architecture | β | Yes |
| 22 industry ontology blueprints | β | Yes |
| Agent autonomy levels (L0-L3) | β | Yes |
| Governance, RBAC, cost controls | β | Yes |
| Multi-tenant operations | β | Yes |
The skills are the appetizer. If you want the full meal β explore the platform.
Get Startedβ
- Star the repo β github.com/frank-luongt/faos-skills-marketplace
- Copy a skill β pick one from your platform folder and drop it in
- Try an agent persona β install a plugin and have a conversation
- Contribute β add your own skills or improve existing ones (Apache 2.0)
930+ skills. 31 agents. 5 platforms. Free.
What's Coming Nextβ
This is Layer 1 of 4 we're open-sourcing from the FAOS stack:
| Layer | What | When |
|---|---|---|
| Layer 1 (today) | Skills Marketplace β 930+ skill instructions | April 2026 |
| Layer 2 | Workflow Templates β 50+ orchestration templates + lean founder toolkit | May 2026 |
| Layer 3 | Industry Blueprints β 22 domain ontologies for AI agents | June 2026 |
| Layer 4 | MCP Server β connect your AI assistant to project intelligence | July 2026 |
Skills tell your AI what. Workflows tell it how. Blueprints give it context. MCP connects it everywhere.
Follow the repos. Star what you use. We're just getting started.
Built by the FAOS team β making humans and organizations permanently relevant in the age of AI.