AI agents call gpu_status to retrieve information from PSKit without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool only reads system telemetry about GPU hardware state. It has no side effects, does not execute commands, does not modify data, and does not create financial obligations. It is a straightforward monitoring/inspection operation that falls squarely into the Read category. The severity is low because even if an AI agent queries this repeatedly, it causes no harm to the system.
From the tool's definition Tool description explicitly states it retrieves GPU status information: 'name, VRAM usage, utilization %, temperature'. The verb 'Get' and lack of any modification, deletion, or command execution language confirms this is a query/retrieval operation.
Attacks that exploit this kind of access
Get NVIDIA GPU status: name, VRAM usage, utilization %, temperature. It is categorised as a Read tool in the PSKit MCP Server, which means it retrieves data without modifying state.
Register the PSKit MCP server in PolicyLayer and add a rule for gpu_status: 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 PSKit. Nothing to install.
gpu_status 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 gpu_status 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 gpu_status. 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.
gpu_status is provided by the PSKit MCP server (nickalus12/pskit). 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 →