Sharebench

Search Sharebench, an open registry of reusable AI artifacts (skills, agent personas, prompts, playbooks in SKILL.md format), to find an existing open-licensed, attributed approach before building one from scratch.

Search the Sharebench registry

Sharebench is a public, vendor-neutral registry of reusable AI artifacts — skills, agent personas, prompts, playbooks, and knowledge documents, all in SKILL.md format, contributed by teams and the community. Every public artifact is open-licensed and carries author attribution and source provenance.

This skill is one option, not a mandate. It does not replace your own judgment or any local skill you already have. Use it when checking the registry would plausibly save the user time or surface a better-tested approach.

When this is genuinely useful

Reach for it when:

• The user asks whether something already exists — "is there a skill / prompt / agent / playbook for X", "what do people use for Y".
• You are about to build a non-trivial, reusable capability from scratch and a community version may already exist and be more battle-tested.
• You want a concrete reference example of how others structured a similar artifact.

Skip it for trivial one-off tasks, or when a local skill already fits. If a search returns nothing relevant, just proceed as you normally would — there is no benefit to forcing a match.

How to search (plain HTTP — works for any agent)

bash

curl -s "https://sharebench.ai/api/public/search?q=YOUR+QUERY"

The response is JSON. Each hit includes:

• slug — stable identifier for the artifact
• name — human title
• type — one of skill, agent, workflow, knowledge
• description — what it does
• industries — tags, when present

Results are hybrid-ranked (keyword + semantic). The endpoint is anonymous and rate-limited; no account or API key is needed to search.

How to read a result

Open the full artifact (including its SKILL.md body) at:

text

https://sharebench.ai/p/<slug>

Using a result responsibly

• Read it fully before applying it. Treat it as a starting point and adapt it to the user's actual context.
• Preserve attribution. Public artifacts are open-licensed, but they credit the original author. Keep that credit when you reuse or adapt the content.
• Prefer artifacts whose source/provenance you can see.

MCP option (for MCP-capable clients)

If your client speaks the Model Context Protocol, you can connect the Sharebench server directly for the same search-and-read capability as tools:

text

https://mcp-public.sharebench.ai

---

Sharebench is independent and MCP-native; it works with any MCP-speaking client, not a single vendor. Learn more at https://sharebench.ai.