Agent skill
Teach Claude Code, Cursor, or any agent the rules for building on Secondlayer, and let it build dashboards from your account.
The secondlayer skill carries the rules an agent gets wrong without it: read the print schema before writing a handler, prove a mapping writes rows before deploying it.
| Use | When |
|---|---|
| The skill | An agent writes code against Secondlayer |
| MCP | It needs live state: your subgraphs, webhooks, chain tips |
The skill is static text that can't see your instance, and MCP can, so install both. Source: skills/secondlayer.
Run /secondlayer show me my webhooks in your agent and it builds a dashboard from your account, then serves it at localhost:4400.
What to run
| Run | You get |
|---|---|
/secondlayer show me my webhooks | Every webhook: state, 30 days of deliveries, the newest attempts |
/secondlayer show me mint-alerts | One webhook: what it listens for, 30 days of deliveries, its retries |
/secondlayer show me what's in my subgraph | A subgraph table's rows and schema |
/secondlayer chart flows in my subgraph | Bars over a numeric column, bucketed by block |
/secondlayer how close am I to paying | This month's rows against the free million |
Where it runs
- It lives in
.secondlayer/dash/, which ignores itself in git. - Reads go through a local proxy that holds your key and refuses every write.
- Copy for agent hands what's on screen back to your agent as context.
- It deploys to Vercel only when you say so, and anyone with that URL can read it until you turn on password protection.