Skip to main content

​
Agent Skill File

If you’re building an AI agent (or configuring one) that connects to BaseLayer, you can use our published SKILL.md file to give it everything it needs to know. The skill file is hosted at: 
https://baselayer.id/SKILL.md
Point your agent framework at this URL, or copy the contents into your agent’s skill/tool configuration.

​
What’s in the skill file

The SKILL.md contains:
  • Connection methods for remote (OAuth), local (stdio), and HTTP (API key) setups
  • All seven MCP tools with parameter signatures and return types
  • Usage guidance telling the agent when to call each tool
  • Examples showing common workflows (recall, remember, inspect)
  • First-time setup checklist for users who haven’t connected yet

​
Why this matters

Most AI agents don’t know BaseLayer exists until you tell them. The skill file acts as a self-contained instruction set. Once an agent reads it, it knows how to connect, which tools are available, and when to use them. This is especially useful for:
  • Agent frameworks like OpenClaw, LangGraph, or CrewAI that support skill/tool discovery
  • Custom agents you’re building that need persistent memory
  • System prompts where you want to include BaseLayer instructions inline

​
Using it with OpenClaw

OpenClaw agents can reference the skill file directly. Add it to your agent’s skill list and it will automatically know how to use BaseLayer for recall and memory.

​
Using it in a system prompt

If your agent framework doesn’t support skill files natively, you can fetch the contents and include them in your system prompt: 
You have access to a BaseLayer knowledge vault via MCP.
Connection: https://api.baselayer.id/mcp (OAuth 2.1)

Available tools: recall, remember, relate, inspect, status, recent, forget

When the user mentions a person, project, or past decision, call recall() first.
When the user states a fact worth preserving, call remember().
For the full version with examples and parameter details, see the hosted skill file.
The skill file encourages agents to call recall liberally. Each result is only about 120 tokens, so it’s cheap to search often and expensive to miss relevant context.

​
Keeping it updated

The skill file at baselayer.id/SKILL.md is always the latest version. If you’ve copied it into your agent config, check back periodically for updates. For the full product context (architecture, all integrations, security model), see llms.txt.