Run a DuckDB SQL query across memories. Use memory names as table names. Example: SELECT * FROM user_preferences WHERE key = \
AI agents invoke query_memory to trigger actions in Structured-sh. 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.
This tool executes arbitrary DuckDB SQL queries, making it Execute category. While the example shows a SELECT, arbitrary SQL can include DELETE, DROP, UPDATE, or other destructive statements. The blast radius is high because a misused query could corrupt or destroy persistent memory state across sessions and tools, and the open-ended SQL execution surface is significant.
From the tool's definition "Run a DuckDB SQL query across memories" — executes arbitrary SQL, which can include destructive or read operations depending on the query
Attacks that exploit this kind of access
Run a DuckDB SQL query across memories. Use memory names as table names. Example: SELECT * FROM user_preferences WHERE key = \. It is categorised as a Execute tool in the Structured-sh MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Structured-sh MCP server in PolicyLayer and add a rule for query_memory: 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 Structured-sh. Nothing to install.
query_memory 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 query_memory 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_memory. 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_memory is provided by the Structured-sh MCP server (structured-sh/structured). 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 →