Search through documentation using semantic similarity matching. Returns relevant document chunks with context.
AI agents call search_docs to retrieve information from Embeddings Searcher without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool reads and queries documentation without side effects. It matches the Read category pattern: it retrieves data based on semantic similarity matching. No write, execute, destructive, or financial operations are possible. The low severity reflects that misuse would only expose existing documentation content.
From the tool's definition Tool 'search_docs' performs 'semantic search' and 'Returns relevant document chunks' — retrieval operations with no modification, deletion, or execution capabilities.
Attacks that exploit this kind of access
Search through documentation using semantic similarity matching. Returns relevant document chunks with context. It is categorised as a Read tool in the Embeddings Searcher MCP Server, which means it retrieves data without modifying state.
Register the Embeddings Searcher MCP server in PolicyLayer and add a rule for search_docs: 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 Embeddings Searcher. Nothing to install.
search_docs 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_docs 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_docs. 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_docs is provided by the Embeddings Searcher MCP server (thypon/kb). 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 →