Skills Marketplace Readiness

Workflow

1. Confirm every skill is an independent folder with a matching SKILL.md frontmatter name and a specific activation-focused description.
2. Keep each skill narrow, executable, and short enough for progressive disclosure. Move long background material to references only when needed.
3. Add install and discovery surfaces:
   • README install command: npx skills@latest add owner/repo
   • skills.sh badge: [![skills.sh](https://skills.sh/b/owner/repo)](https://skills.sh/owner/repo)
   • .claude-plugin/plugin.json with all published skill paths.
   • scripts/list-skills.sh and scripts/link-skills.sh for local fallback.
4. Validate that all plugin paths resolve and each listed skill has a SKILL.md.
5. Confirm safety-relevant skills — those that can drive autonomous checkout, payment, support actions, or other agent-executed changes — have a behavioral eval (fixtures plus a pass/fail runner) wired into CI. A skill that only has prose guardrails and no eval coverage is not marketplace-ready.
6. Position the repo around searchable intent, not internal naming: use words buyers and agents search for, such as ecommerce, agentic commerce, AEO, GEO, SEO, agent readiness, checkout, ACP, UCP, MCP, and AI crawler readiness.
7. Submit or seed marketplace visibility by installing with the public CLI, sharing the GitHub URL, adding the badge, and linking the repo from website, docs, social profiles, and relevant skill directories.

Guardrails