A personal collection of Agent Skills for AI coding agents.
# Install all skills
npx skills add shaunburdick/skills
# List available skills without installing
npx skills add shaunburdick/skills --list
# Install a specific skill
npx skills add shaunburdick/skills --skill spec-driven-development
# Install globally (available across all projects)
npx skills add shaunburdick/skills -g
# Install to specific agents only
npx skills add shaunburdick/skills -a claude-code -a opencodeStructured spec-driven development workflow: clarify requirements, write specs, plan architecture, break into tasks, then implement. Use when starting a new feature, building from scratch, or when requirements are unclear.
Practical setup and usage guide for the spec-kit CLI — covers installation, project initialization for Claude Code and OpenCode, the /speckit.* slash commands, helper scripts, and detecting current workflow state. Pairs with the spec-driven-development skill: this one covers the tooling, that one covers the methodology.
Enforces safe git practices for AI coding agents. Defines branch protection rules (never commit to main/master/develop), commit policies, amend rules, and recommended agent permissions. Use when configuring any agent that writes code and commits to git repositories.
Standardizes Generated-By attribution for AI-assisted commits, pull requests, comments, issues, and optional source disclosures. Use when agent-generated content will appear under a human account.
Enforces non-negotiable code quality standards: zero lint suppressions (eslint-disable, @ts-ignore, etc.), strict TypeScript type safety (no any), and a mandatory pre-commit verification protocol. Use when any agent writes, edits, or reviews code.
Quantifies where a repository's lines actually go, so documentation and comment bloat becomes a number in a diff rather than a surprise a year later. Reports comment-vs-code share, test lines with test and assertion counts, spec and doc prose (both gathered and scattered), and agent instruction surface. Ships a dependency-free script with calibrated threshold warnings and an opposing-direction ratchet whose bands mean reducing documentation has to be a deliberate baseline change rather than something achieved by deletion. Composition only — no metric here is a quality score.
Install @shaunburdick's personal style configuration: shared .editorconfig for any project, plus a choice of eslint-config-shaunburdick or biome-config-shaunburdick for JavaScript/TypeScript projects, with a hand-off to those package skills for detailed setup and debugging.
Author secure, maintainable GitHub Actions workflows with security-first defaults. Every third-party action is pinned to a full commit SHA (never a mutable tag), permissions are least-privilege, and dependencies stay current with Dependabot. Covers supply-chain hardening, reusable workflows vs. composite actions, OIDC auth, and the OpenSSF Scorecard. Use for any CI/CD or workflow automation task.
Design, build, test, and integrate standards-based custom elements and framework-agnostic Web Components. Covers lifecycle timing, Light DOM and Shadow DOM composition, accessibility, registration, SSR boundaries, CSS integration, React interoperability, Vitest/Playwright testing, and package design. Use when building web components, custom elements, or any work involving customElements.define, HTMLElement, slots, or attachShadow.
Each skill lives in skills/<skill-name>/ and contains:
skills/
└── skill-name/
├── SKILL.md # Required: YAML frontmatter + agent instructions
├── scripts/ # Optional: helper scripts the agent can run
├── references/ # Optional: supplementary documentation
└── assets/ # Optional: templates, schemas, static resources
---
name: skill-name
description: What this skill does and when to use it. Be specific — agents use this to decide when to activate the skill.
license: MIT
metadata:
author: shaunburdick
version: "1.0.0"
---
# Skill Title
Instructions for the agent...See the Agent Skills specification for the full format reference.
Feel free to fork this repo and add your own skills. The skills/ directory is the canonical location the skills CLI searches.
MIT