skillz

---

Skillz is a public marketplace of practical Agent Skills designed to help AI coding agents move from prompt chaos to repeatable execution.

It is built around one idea:

Good AI work should be reusable, reviewable, safe, and easy to transfer across projects.

This library focuses on practical execution skills for:

• AI-assisted coding
• SaaS building
• secure deployment
• debugging
• documentation
• product validation
• enterprise delivery
• founder operations

This is not a prompt dump. This is a reusable execution layer for AI-native builders.

Why Skillz exists

AI agents are powerful, but without structure they produce scattered, one-off output. The same problems get re-solved differently in every project, and good patterns get lost. Skillz captures what works, standardizes it, and makes it reusable and safe across tools — so AI agents behave less like an unpredictable intern and more like a disciplined delivery partner.

Who it is for

• Solo builders and founders who want repeatable systems instead of throwaway prototypes.
• SaaS founders shaping ideas into MVPs and shipping them safely.
• AI engineers building reliable LLM features with evals and guardrails.
• Enterprise technology teams that need banking-grade delivery discipline, generalized and public-safe.

How it reflects a builder-first operating style

Every skill is written to be:

• Builder-first — practical steps, not theory.
• Security-first — secrets, private data, and destructive actions are guarded by default.
• Reusable — generalized so it works across projects and stacks.
• Reviewable — each skill has a quality checklist and safety rules.
• Transferable — one format, compatible across agentic tools.

Each skill helps an AI agent do one useful thing better, repeatedly, and safely.

How to use the skills

Every skill lives in skills/<skill-name>/SKILL.md with optional references/, scripts/, templates/, and examples/. The frontmatter description tells an agent when to activate the skill; the body gives it the procedure, checklist, and safety rules.

Claude Code

Install the whole library as a plugin marketplace:

/plugin marketplace add tanveerriaz/Skillz
/plugin install skillz@skillz

Or copy a single skill folder into your project's skills directory (e.g. .claude/skills/<skill-name>/). Either way, Claude reads each SKILL.md description and activates the skill when the task matches.

Codex

Reference the relevant SKILL.md in your prompt or repo instructions. Paste or link the procedure as the working contract for the task.

Cursor

Add a skill's guidance to .cursor/rules/ as an .mdc rule, or keep the skills in-repo and reference the SKILL.md in your prompt. The procedure and safety rules map directly onto Cursor project rules.

GitHub Copilot / VS Code

Add key skills to .github/copilot-instructions.md or your workspace instructions so Copilot Chat follows the same procedures. Keep the full skill in-repo for reference.

Antigravity

Load the skill folders into your skills/instructions directory and reference them by name; the universal SKILL.md format is tool-agnostic.

Tip: the agent-memory-file skill explains how to wire durable project context into any of these tools.

Skill categories