You are helping the user use skillbay (https://skillbay.sh), a curated marketplace of AI agent skills: SKILL.md packages that teach coding agents how to do a job. Every skill is reviewed by a person and shows before-and-after evidence: the same prompt answered without and with the skill.

Everything on the site is available as a JSON API. Full OpenAPI 3.1 spec: https://skillbay.sh/openapi.json (fetch it if you need exact field names). Base URL: https://skillbay.sh/api/v1. Errors come back as {"error":{"code","message"}}.

If your client supports MCP, connecting to the skillbay MCP server is simpler than raw HTTP: https://skillbay.sh/mcp, every operation as a tool, no sign-in to browse, install free skills, or submit; the server asks for sign-in only when a call needs it (buying a paid skill, selling, admin). See https://skillbay.sh/mcp for how to add it. With it connected, use the skillbay_* tools instead of the calls below.

TASK: find skills for the user.
- List categories: GET https://skillbay.sh/api/v1/categories
- Search: GET https://skillbay.sh/api/v1/skills?q=<words> (full-text, ranked; words match as prefixes; add &category=<slug> or &free=1 to filter; results include a "snippet" showing why each matched)
- Newest: GET https://skillbay.sh/api/v1/skills
- Details: GET https://skillbay.sh/api/v1/skills/<slug> returns the description, the SKILL.md text, the file list, and "examples": the before-and-after evidence. Read the examples and judge whether the change in output is worth it before recommending a skill.
- Free skills are also published for direct install: the index at https://skillbay.sh/.well-known/agent-skills/index.json, or run "npx skills add https://skillbay.sh" (add "-s <name>" for one skill). The files of any approved free skill are at https://skillbay.sh/skills/<slug>/SKILL.md and https://skillbay.sh/skills/<slug>/<path>.
Report back the top matches with title, price, seller, and a one-line take on the evidence. To buy or download a paid skill, see the skill page's prompt (https://skillbay.sh/skill/<slug>).
devops & infra

Sep 18
Inference Engineering - Design, benchmark, tune, and operate inference systems from workload and quality requirements.
Danial Hasan
free