My personal collection of agent skills for Claude Code (and other agentskills.io-compatible agents)
npx skills add https://github.com/rogeriochaves/skills --skill orchestrateقم بتثبيت هذه المهارة باستخدام واجهة سطر الأوامر (CLI) وابدأ في استخدام سير عمل SKILL.md في مساحة عملك.
My personal collection of agent skills for Claude Code and other skills-compatible agents.
Globally (available in every project):
npx skills add rogeriochaves/skills -g
Into the current project only:
npx skills add rogeriochaves/skills
A single skill:
npx skills add rogeriochaves/skills --skill orchestrate -g
List what's in here without installing:
npx skills add rogeriochaves/skills --list
| Skill | Description |
|---|---|
orchestrate |
Disciplined delivery loop: BDD specs → real integration tests (no mocks) → PR + CI/CodeRabbit babysitting → mandatory end-user QA via computer-use or CLI dogfooding before anything is called done. |
browser-qa |
Drive a real browser to QA a feature end-to-end. Picks the right stack (Playwright MCP / Claude-in-Chrome / computer-use), covers failure modes, captures PR screenshots. The QA stage of orchestrate. |
drive-pr |
Babysit a PR to mergeable state — address CodeRabbit/reviewer comments, fix CI failures (even pre-existing ones), push fixes, then loop every 4.5 min until two consecutive clean checks. |
Every skill lives at skills/<name>/SKILL.md and follows the Agent Skills specification — YAML frontmatter (name, description) followed by the instructions the agent loads when the skill activates.
For my own setup I don't want npx skills add because it copies files to ~/.agents/skills/ — edits to this repo wouldn't flow through until I reinstall. Instead I symlink each skill directory straight into ~/.claude/skills/ so changes here are picked up by every Claude Code session immediately:
git clone [email protected]:rogeriochaves/skills.git ~/Projects/skills
ln -s ~/Projects/skills/skills/orchestrate ~/.claude/skills/orchestrate
Repeat the ln -s line for each new skill. Verify with ls -la ~/.claude/skills/.