Stop backend to clear variables and free memory. Use when you want a clean slate or the backend is misbehaving. The backend will restart automatically when needed.
AI agents invoke stop to trigger actions in Scicompute. 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 a backend control operation that interrupts running processes and clears state. While not destructive (the data/variables are temporary and cleared by design, not permanently deleted from external storage) and not immediately harmful, it is an active operation that modifies system state and could disrupt ongoing computations if misused by an agent.
From the tool's definition Tool description states it 'Stop backend to clear variables and free memory' and 'The backend will restart automatically when needed.' This actively triggers an operation (backend shutdown/restart) whose effects depend on the backend state and subsequent…
Attacks that exploit this kind of access
Stop backend to clear variables and free memory. Use when you want a clean slate or the backend is misbehaving. The backend will restart automatically when needed. It is categorised as a Execute tool in the Scicompute MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Scicompute MCP server in PolicyLayer and add a rule for stop: 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 Scicompute. Nothing to install.
stop 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 stop 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 stop. 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.
stop is provided by the Scicompute MCP server (sanshanjianke/scicompute-mcp). 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 →