Hybrid search combining semantic search + BM25 keyword search with cross-encoder reranking. Read-only. No side effects. Args: query: Search query text (1–3 keywords recommended; phrase queries also work) max_results: Maximum number of results (default: 5, max: 20) category: Optional category filt...
AI agents call search_knowledge to retrieve information from Knowledge Rag without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This is a pure retrieval tool that queries a local RAG knowledge base and returns search results. It accepts only read-oriented parameters (search terms, result limits, filters) and produces no modifications to data or side effects. The 'read-only' designation and query-only nature classify it as Read with minimal risk.
From the tool's definition Tool explicitly stated as 'Read-only. No side effects.' Performs 'Hybrid search combining semantic search + BM25 keyword search with cross-encoder reranking' with parameters for query, max_results, category filter, and hybrid_alpha.
Documented attack patterns abuse exactly the kind of access search_knowledge gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Knowledge Rag, and nothing reaches the server without passing your rules. This is the rule we recommend for search_knowledge:
{
"version": "1",
"default": "deny",
"tools": {
"search_knowledge": {}
}
}search_knowledge is read-only, so it stays allowed — but everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Hybrid search combining semantic search + BM25 keyword search with cross-encoder reranking. Read-only. No side effects. Args: query: Search query text (1–3 keywords recommended; phrase queries also work) max_results: Maximum number of results (default: 5, max: 20) category: Optional category filter — one of: security, ctf, logscale, development, general, redteam, blueteam. Call list_categories() first to see available categories and counts. hybrid_alpha: Balance between semantic and keyword search. 0.0 = keyword-only (best for exact technical terms like CVE IDs or tool names), 0.3 = balanced default, 1.0 = semantic-only (best for conceptual or natural-language queries). Returns: JSON string with results including content chunks, source filepath, relevance score, and search method used. Returns chunks, not full document content. Usage: Primary search tool — use for any topic or keyword lookup. Prefer search_similar() when you already have a reference document and want more like it. Prefer get_document() when you already know the exact filepath and need the full content. It is categorised as a Read tool in the Knowledge Rag MCP Server, which means it retrieves data without modifying state.
Register the Knowledge Rag MCP server in PolicyLayer and add a rule for search_knowledge: 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 Knowledge Rag. Nothing to install.
search_knowledge 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_knowledge 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_knowledge. 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_knowledge is provided by the Knowledge Rag MCP server (lyonzin/knowledge-rag). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Knowledge Rag, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
12 Knowledge Rag tools catalogued and risk-classified — across an index of 43,000+ MCP servers.