Low Risk

pg_query

Execute a read-only SQL query against the PostgreSQL database. Args: query: The SQL query to execute (must be read-only) conn_id: Connection ID previously obtained from the connect tool params: Parameters for the query (optional) Returns: Query results as a list of dictionaries

How to control pg_query ↓

AI agents call pg_query to retrieve information from Pg without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.

Low Risk

pg_query is restricted to read-only operations against a PostgreSQL database. It retrieves and returns data without creating, modifying, or deleting any records. The explicit read-only constraint and parameter validation in the description confirm this tool cannot perform writes, destructive operations, or execute arbitrary code.

From the tool's definition Tool description explicitly states 'Execute a read-only SQL query' and 'must be read-only'. The tool only returns query results as dictionaries with no data modification capability.

Documented attack patterns abuse exactly the kind of access pg_query gives an agent:

PolicyLayer is an MCP gateway — it sits between your AI agents and Pg, and nothing reaches the server without passing your rules. This is the rule we recommend for pg_query:

policy.json
{
  "version": "1",
  "default": "deny",
  "tools": {
    "pg_query": {}
  }
}

pg_query is read-only, so it stays allowed — but everything else on the server is denied unless you say otherwise.

  1. Create a free account and register Pg — nothing to install.
  2. Add this policy — paste it, or build it visually.
  3. Point your MCP client (Claude, Cursor, anything) at your gateway URL.
CAP THIS TOOL →

Free to start. No card required.

Go deeper

What does the pg_query tool do? +

Execute a read-only SQL query against the PostgreSQL database. Args: query: The SQL query to execute (must be read-only) conn_id: Connection ID previously obtained from the connect tool params: Parameters for the query (optional) Returns: Query results as a list of dictionaries. It is categorised as a Read tool in the Pg MCP Server, which means it retrieves data without modifying state.

How do I enforce a policy on pg_query? +

Register the Pg MCP server in PolicyLayer and add a rule for pg_query: 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 Pg. Nothing to install.

What risk level is pg_query? +

pg_query is a Read tool with low risk. Read-only tools are generally safe to allow by default.

Can I rate-limit pg_query? +

Yes. Add a rate_limit block to the pg_query 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.

How do I block pg_query completely? +

Set action: deny in the PolicyLayer policy for pg_query. 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.

What MCP server provides pg_query? +

pg_query is provided by the Pg MCP server (stuzero/pg-mcp-server). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.

Enforce policy on every Pg tool call.

Deterministic rules across all 5 Pg tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.

Free to start. No card required.

5 Pg tools catalogued and risk-classified — across an index of 42,500+ MCP servers.

// GET IN TOUCH

Have a question or want to learn more? Send us a message.

Message sent.

We'll get back to you soon.