AI agents invoke mysql_query to trigger actions in Mysql. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call — builds kicked off, notifications sent, workflows started.
Because the tool accepts and runs arbitrary MySQL queries, it spans Read through Destructive. Under the most-severe-applicable rule, it must be classified as Destructive (DROP/DELETE are possible), but since it also covers non-destructive queries, Execute is the correct category when the effect depends on arguments.
From the tool's definition 'Execute any MySQL query' — the tool runs arbitrary SQL, which can include SELECT, INSERT, UPDATE, DELETE, DROP, or any other statement.
Attacks that exploit this kind of access
Execute any MySQL query and return results. It is categorised as a Execute tool in the Mysql MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Mysql MCP server in PolicyLayer and add a rule for mysql_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 Mysql. Nothing to install.
mysql_query is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.
Yes. Add a rate_limit block to the mysql_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.
Set action: deny in the PolicyLayer policy for mysql_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.
mysql_query is provided by the Mysql MCP server (myheisenberg/mysql-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 →