AI agents invoke agent_cancel to trigger actions in Musubix. 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 cancels running agent tasks and interrupts subtasks. It triggers an external operation (cancellation/interruption) against live agent processes. While it doesn't delete data, it actively modifies the execution state of running tasks, which could have non-trivial side effects depending on what those agents were doing. Execute is the most appropriate category as it triggers an external operational change.
From the tool's definition 「実行中のエージェントタスクをキャンセルします」「まだ開始されていないサブタスクはスキップされ、実行中のタスクは可能な限り中断されます」
Attacks that exploit this kind of access
実行中のエージェントタスクをキャンセルします。 まだ開始されていないサブタスクはスキップされ、実行中のタスクは可能な限り中断されます。. It is categorised as a Execute tool in the Musubix MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Musubix MCP server in PolicyLayer and add a rule for agent_cancel: 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 Musubix. Nothing to install.
agent_cancel 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 agent_cancel 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 agent_cancel. 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.
agent_cancel is provided by the Musubix MCP server (@nahisaho/musubix-mcp-server). 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 →