Undo the last edit in the focused widget (ctrl+z).
AI agents invoke undo to trigger actions in Hermes Computer Use. 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 OS-level keyboard action (Ctrl+Z) in a real browser, which executes a command that modifies the state of a UI widget by reverting its last edit. While it superficially reverses a change, it is an active browser/OS operation with side effects dependent on the current context—fitting the Execute category. It could inadvertently restore sensitive or unintended content in the focused element.
From the tool's definition Undo the last edit in the focused widget (ctrl+z)
Attacks that exploit this kind of access
Undo the last edit in the focused widget (ctrl+z). It is categorised as a Execute tool in the Hermes Computer Use MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Hermes Computer Use MCP server in PolicyLayer and add a rule for undo: 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 Hermes Computer Use. Nothing to install.
undo 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 undo 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 undo. 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.
undo is provided by the Hermes Computer Use MCP server (noah3521/hermes-computer-use). 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 →