Agent Skills Repository
Sharquille/agent-skills-repository
Registry seguro de skills auditadas com deploy automático para agentes IA
Ver no GitHubRepositório curado e auditado de skills production-grade para Claude Code, Claude Desktop e Gemini CLI. Oferece um registro centralizado com versionamento, tooling de deploy automatizado e auditorias de segurança integradas para gerenciar capacidades avançadas de agentes de IA.
Instalação
npx convez-ai Sharquille/agent-skills-repository1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Centralizar e versionare skills customizadas para Claude e Gemini
- Auditar e integrar skills de terceiros com garantia de segurança
- Automatizar deployment de capacidades em múltiplos ambientes de IA
Veredito

Solução pragmática para quem trabalha com múltiplos agentes IA e precisa gerenciar skills em escala. O foco em segurança (auditoria obrigatória) e estrutura clara de versionamento é um diferencial real, embora exija familiaridade com CLI e gestão de repositórios.
README
A tradução para o português deste README ainda está na fila.
Agent Skills Repository
A curated, secure-by-default registry of high-signal skills for AI agents like Claude Code, Claude Desktop, and Gemini CLI.
This repository serves as a personal collection of self-authored and audited third-party skills, providing a central place to manage, version, and deploy advanced capabilities to your local coding assistants.
🚀 Getting Started
Quick Install (Claude Only)
For a basic setup that symlinks the entire skills/ directory to ~/.claude/skills:
./install.sh
Full Deployment (Claude & Gemini CLI)
To deploy skills to both Claude and Gemini CLI (handling Gemini's specific directory nesting requirements):
./scripts/deploy.sh
See deploy-agent-skills for more details.
📂 Repository Structure
skills/: The core registry, categorized for easy discovery.design/: UI/UX, site architecture, and visual planning.engineering/: Security, network forensics, cloud architecture, and development workflows.productivity/: Brainstorming, resume review, and knowledge management.
scripts/: Automation for deployment and maintenance.REGISTRY.md: The "Source of Truth" table tracking every skill, its author, and provenance.NOTICE.md: Detailed attribution and licensing information.
🛡️ Security & Auditing
Every third-party skill in this repository has been vetted before inclusion. We recommend using the built-in vet-skill to audit any new skills before adding them to your registry.
How to add a new skill
- Download the skill files to a temporary directory.
- Run the audit:
skills/engineering/vet-skill/scripts/audit.sh /tmp/new-skill. - If clean, move to the appropriate category in
skills/. - Update
REGISTRY.md. - Run
./scripts/deploy.shto update your global agent environment.
📜 License & Attribution
This repository uses a mixed-licensing model:
- Original curation, tooling, and self-authored skills are under the MIT License.
- Third-party skills retain their original licenses (e.g., Apache-2.0, MIT).
See NOTICE.md for full details and proper attribution for all contributors.