// MCP TOOL REFERENCE

AGENT RECALL TOOLS

6 tools from the Agent Recall MCP Server, categorised by risk level.

READ TOOLS

check Record what you think the human wants. Returns predictive warnings based on past correction patterns. digest Context cache — store or recall pre-computed analysis results. Use 'store' after expensive operations (subagent explorations, code analysis) to cac... 2/5 recall Search all memory stores at once. Returns ranked results from palace, journal, and insights. 2/5

WRITE TOOLS

remember Save a memory. Auto-classifies and routes to the right store (journal, palace, knowledge, or awareness). 2/5 session_end Save session summary, insights, and trajectory. Writes journal, updates awareness, consolidates to palace. 2/5

EXECUTE TOOLS

session_start Load project context for a new session. Returns identity, insights, active rooms, and recent activity in one call. 3/5

// FAQ

How many tools does the Agent Recall MCP server have? +

The Agent Recall MCP server exposes 6 tools across 3 categories: Read, Write, Execute.

How do I enforce policies on Agent Recall tools? +

Use Intercept, the open-source MCP proxy. Write YAML rules for each tool — rate limits, argument validation, or deny rules — then run Intercept in front of the Agent Recall server.

What risk categories do Agent Recall tools fall into? +

Agent Recall tools are categorised as Read (3), Write (2), Execute (1). Each category has a recommended default policy.

Enforce policies on Agent Recall

Open source. One binary. Zero dependencies.

npx -y @policylayer/intercept

github.com/policylayer/intercept →