Semantic search via embedding cosine similarity. Searches all vaults by default.
AI agents call search_semantic to retrieve information from Vault Memory without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves and queries data (searches vaults) with no side effects on the underlying knowledge base. It is a pure read operation analogous to search or query functions. The low severity reflects minimal blast radius—semantic search results cannot damage or compromise the system, and misuse would only return irrelevant query results rather than expose secrets or cause harm.
From the tool's definition Tool performs 'semantic search via embedding cosine similarity' with no modification capability. Sibling tools include destructive operations (delete_note) and write operations (compile_brief), but search_semantic is clearly a retrieval function that…
Attacks that exploit this kind of access
Semantic search via embedding cosine similarity. Searches all vaults by default. It is categorised as a Read tool in the Vault Memory MCP Server, which means it retrieves data without modifying state.
Register the Vault Memory MCP server in PolicyLayer and add a rule for search_semantic: 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 Vault Memory. Nothing to install.
search_semantic 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 search_semantic 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 search_semantic. 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.
search_semantic is provided by the Vault Memory MCP server (owrede/vault-memory). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →