AI agents call deep_search to retrieve information from Astrolabe without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool belongs to a documentation indexing/search server (astrolabe-mcp) whose purpose is unified search and navigation. Despite the empty description lowering confidence slightly, the contextual placement with other Read operations (search_docs, list_docs, read_doc) and the semantic meaning of 'deep_search' (searching, not modifying) indicates this is a Read operation.
From the tool's definition Tool is listed as a sibling alongside search_docs, list_docs, read_doc, and get_card — all data retrieval operations typical of a documentation search server. The name 'deep_search' indicates comprehensive searching without modification.
Attacks that exploit this kind of access
deep_search. It is categorised as a Read tool in the Astrolabe MCP Server, which means it retrieves data without modifying state.
Register the Astrolabe MCP server in PolicyLayer and add a rule for deep_search: 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 Astrolabe. Nothing to install.
deep_search 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 deep_search 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 deep_search. 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.
deep_search is provided by the Astrolabe MCP server (zebrr/astrolabe-mcp). 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 →