AI agents invoke stop_playback to trigger actions in Vmd. 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 triggers an external operation (stopping VMD's animation playback) whose effects depend on the current state of the application. While not destructive or creating side effects on data, it executes a control action in VMD. The severity is medium because misuse would interrupt user workflows but causes no permanent data loss or financial impact.
From the tool's definition Tool name 'stop_playback' and description 'Stop any currently running trajectory animation in VMD' indicates execution of a control command that halts an active process in the VMD molecular dynamics software.
Attacks that exploit this kind of access
Stop any currently running trajectory animation in VMD. It is categorised as a Execute tool in the Vmd MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Vmd MCP server in PolicyLayer and add a rule for stop_playback: 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 Vmd. Nothing to install.
stop_playback 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_playback 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_playback. 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_playback is provided by the Vmd MCP server (omararias-gaguancela/vmd-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 →