A curated marketplace of reusable AI Agent Skills for Infrastructure, DevOps, and Automation. Each Skill is a standardized, battle-tested knowledge module that any AI Agent can use. Skills follow the Agent Skills standard — the open format supported by 30+ agent products.
Every Skill can be consumed in three ways:
| Method | How it works |
|---|---|
| Claude Code | Install via plugin marketplace, then ask the agent to run the skill |
| CLI | Run bash scripts directly from the terminal |
| GitHub Actions | Add to CI/CD pipelines via action.yml |
- bash 4.0+ — required for CLI and GitHub Actions (scripts use associative arrays)
- macOS ships bash 3.2; upgrade with
brew install bash - Linux (Ubuntu, Debian, etc.) ships bash 5.x — no action needed
- CI (GitHub Actions
ubuntu-latest) — no action needed
- macOS ships bash 3.2; upgrade with
grep,find,file,wc— standard POSIX tools (pre-installed everywhere)- Optional:
gitleaks,shellcheck,trivy,jq— enhanced checks, gracefully skipped if missing
/plugin marketplace add PSDN-AI/nexus-skillsgit clone https://github.com/PSDN-AI/nexus-skills.git- uses: PSDN-AI/nexus-skills/skills/<skill-name>@main| Skill | Category | Description |
|---|---|---|
| prd-decompose | Product Engineering | Decompose a PRD into domain-specific specs for AI Agent consumption |
| spec-plan | Product Engineering | Convert a domain spec into an executable task graph with dependency ordering and parallelization |
| agent-launcher | Product Engineering | Execute a task graph with isolated sub-agents, dependency enforcement, and run reports |
| gha-create | CI/CD & DevOps | Generate GitHub Actions workflows with elite security and efficiency practices |
| repo-audit | Security & Compliance | Scan for secrets, quality issues, and compliance problems before going public |
See CONTRIBUTING.md for how to add a new Skill.