Give your AI agent the same on-chain verdict LENS shows you, a token or dev check it can run in the middle of a conversation
A Skill is a short note an agent reads to learn how and when to use a tool. The LENS Skill tells an agent to reach for LENS the moment a user pastes a Base contract address, names a ticker, or asks if a token or dev is safe, then how to call it and how to read the verdict.
# lens-onchain-intel · SKILL.md when to use user pastes a Base contract address and asks if it is safe user names a Base token ticker and wants a risk read user asks about a crypto X account, has the dev sold, is it a rug how to call POST https://lens-liard.vercel.app/api/agent { "q": "<contract address, $ticker, or @handle>" } read the verdict CLEAR no hard red flags CAUTION one or two red lines triggered STOP three or more red lines triggeredopen SKILL.md →
No connector needed to start. Any agent can POST a contract address, a ticker, or a handle to the LENS endpoint and get a full breakdown back, verdict included.
// one call, one answer POST https://lens-liard.vercel.app/api/agent Content-Type: application/json { "q": "0xb233BDFFD437E60fA451F62c6c09D3804d285Ba3" } // response { "ok": true, "type": "ca", "answer": "...breakdown, ends with Verdict: STOP and the red lines...", "data": { ...raw fields... } }
Show the user the answer field as given. For a token it ends with the verdict and the exact red lines that triggered, never invent or soften them.
A native MCP endpoint so an agent can discover LENS tools, quote a call, and run it without any custom glue. Same data and the same verdict, exposed as proper agent tools.
# add LENS to your agent
claude mcp add --transport http \
lens https://lens-liard.vercel.app/api/mcp