Read back a value previously stored via memory_store, by exact (namespace, key) — lossless, includes metadata. Use when native Read is wrong because the value is not a file (it lives in the .swarm/memory.db SQLite store) AND you know the exact key. For semantic lookup by meaning, use memory_search.
AI agents call memory_retrieve to retrieve information from Claude Flow without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool performs a pure retrieval operation on stored data. It queries an internal SQLite memory store to fetch previously-stored values by key, returning them 'lossless, includes metadata.' There are no side effects, no data modification, no code execution, and no irreversible actions.
From the tool's definition Tool explicitly described as 'Read back a value previously stored' and 'Use when native Read is wrong' — it retrieves data from memory storage without modification.
Attacks that exploit this kind of access
Read back a value previously stored via memory_store, by exact (namespace, key) — lossless, includes metadata. Use when native Read is wrong because the value is not a file (it lives in the .swarm/memory.db SQLite store) AND you know the exact key. For semantic lookup by meaning, use memory_search. It is categorised as a Read tool in the Claude Flow MCP Server, which means it retrieves data without modifying state.
Register the Claude Flow MCP server in PolicyLayer and add a rule for memory_retrieve: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Claude Flow. Nothing to install.
memory_retrieve is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the memory_retrieve rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for memory_retrieve. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
memory_retrieve is provided by the Claude Flow MCP server (claude-flow). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.