Query an Amazon Bedrock Knowledge Base using natural language. ## Usage Requirements - You MUST first use the ListKnowledgeBases tool to get valid knowledge base IDs - You can query different knowledge bases or make multiple queries to the same knowledge base ## Query Tips - Use clear, specific...
Accepts freeform code/query input (query)
Part of the Amazon Bedrock Knowledge Base Retrieval MCP Server MCP server. Enforce policies on this tool with Intercept, the open-source MCP proxy.
AI agents call QueryKnowledgeBases to retrieve information from Amazon Bedrock Knowledge Base Retrieval MCP Server without modifying any data. This is common in research, monitoring, and reporting workflows where the agent needs context before taking action. Because read operations don't change state, they are generally safe to allow without restrictions -- but you may still want rate limits to control API costs.
Even though QueryKnowledgeBases only reads data, uncontrolled read access can leak sensitive information or rack up API costs. An agent caught in a retry loop could make thousands of calls per minute. A rate limit gives you a safety net without blocking legitimate use.
Read-only tools are safe to allow by default. No rate limit needed unless you want to control costs.
tools:
QueryKnowledgeBases:
rules:
- action: allowSee the full Amazon Bedrock Knowledge Base Retrieval MCP Server policy for all 2 tools.
Agents calling read-class tools like QueryKnowledgeBases have been implicated in these attack patterns. Read the full case and prevention policy for each:
Other tools in the Read risk category across the catalogue. The same policy patterns (rate-limit, allow) apply to each.
Query an Amazon Bedrock Knowledge Base using natural language. ## Usage Requirements - You MUST first use the ListKnowledgeBases tool to get valid knowledge base IDs - You can query different knowledge bases or make multiple queries to the same knowledge base ## Query Tips - Use clear, specific natural language queries for best results - You can use this tool MULTIPLE TIMES with different queries to gather comprehensive information - Break complex questions into multiple focused queries - Consider querying for factual information and explanations separately ## Tool output format The response contains multiple JSON objects (one per line), each representing a retrieved document with: - content: The text content of the document - location: The source location of the document - score: The relevance score of the document ## Interpretation Best Practices 1. Extract and combine key information from multiple results 2. Consider the source and relevance score when evaluating information 3. Use follow-up queries to clarify ambiguous or incomplete information 4. If the response is not relevant, try a different query, knowledge base, and/or data source 5. After a few attempts, ask the user for clarification or a different query.. It is categorised as a Read tool in the Amazon Bedrock Knowledge Base Retrieval MCP Server MCP Server, which means it retrieves data without modifying state.
Add a rule in your Intercept YAML policy under the tools section for QueryKnowledgeBases. You can allow, deny, rate-limit, or validate arguments. Then run Intercept as a proxy in front of the Amazon Bedrock Knowledge Base Retrieval MCP Server MCP server.
QueryKnowledgeBases 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 QueryKnowledgeBases rule in your Intercept 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 Intercept policy for QueryKnowledgeBases. 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.
QueryKnowledgeBases is provided by the Amazon Bedrock Knowledge Base Retrieval MCP Server MCP server (awslabs.bedrock-kb-retrieval-mcp-server). Intercept sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Deterministic policy on every MCP tool call. Per-identity grants. Full audit log.