Single rich call describing a table: columns (type/nullable/default + null %/distinct ratio from pg_stats), primary key, foreign keys going OUT (this table → others) AND coming IN (others → this table), all indexes (with definitions), table size, row-count estimate, and sample rows. Replaces ~5 s...
AI agents call describe_table to retrieve information from Postgres without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
describe_table exclusively performs introspective queries against PostgreSQL system catalogs (pg_stats, etc.) and returns metadata about table structure and statistics. It has no side effects—no data is created, modified, deleted, or overwritten. The tool serves schema exploration, which is a classic Read operation. The 'sample rows' retrieval is still read-only data retrieval.
From the tool's definition Tool retrieves table metadata: 'columns (type/nullable/default + null %/distinct ratio from pg_stats)', 'primary key', 'foreign keys', 'all indexes (with definitions)', 'table size', 'row-count estimate', and 'sample rows'.
Attacks that exploit this kind of access
Single rich call describing a table: columns (type/nullable/default + null %/distinct ratio from pg_stats), primary key, foreign keys going OUT (this table → others) AND coming IN (others → this table), all indexes (with definitions), table size, row-count estimate, and sample rows. Replaces ~5 separate calls (get_object_details + LIMIT 5 + COUNT(*) + pg_stats). It is categorised as a Read tool in the Postgres MCP Server, which means it retrieves data without modifying state.
Register the Postgres MCP server in PolicyLayer and add a rule for describe_table: 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 Postgres. Nothing to install.
describe_table 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 describe_table 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 describe_table. 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.
describe_table is provided by the Postgres MCP server (teja-sudo/postgres-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 →