AI agents call query_chroma_db to retrieve information from Mcp without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
Query operations on databases are Read category—they retrieve data without modifying, deleting, or executing arbitrary code. The sibling tools (get_docs, setup_chroma_db) reinforce this is a search/retrieval context. Severity is low because vector database queries are non-destructive and their blast radius is limited to information disclosure.
From the tool's definition Tool name 'query_chroma_db' indicates a query operation against ChromaDB, a vector database. The server description confirms this is for 'vector database management' and 'documentation search.' The empty description limits certainty, but querying a vector…
Attacks that exploit this kind of access
query_chroma_db. It is categorised as a Read tool in the Mcp MCP Server, which means it retrieves data without modifying state.
Register the MCP server in PolicyLayer and add a rule for query_chroma_db: 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 Mcp. Nothing to install.
query_chroma_db 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 query_chroma_db 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 query_chroma_db. 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.
query_chroma_db is provided by the MCP server (seonokkim/mcp-server). 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 →