AI agents use handoff_task to create or update resources in Agent Bus — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Agent Bus environment.
This tool writes/creates new data structures (handoff memory records) and conditionally modifies agent assignments. It does not execute arbitrary code, delete data irreversibly, or move money. The blast radius is medium because misuse could redirect critical tasks to wrong agents or create orphaned/duplicate handoffs, causing workflow confusion, but data can be corrected or cleared.
From the tool's definition Tool creates a 'pinned handoff memory' for a task; the verb 'create' combined with 'memory' indicates generation/modification of persistent data state.
Attacks that exploit this kind of access
Create a pinned handoff memory for a task and optionally assign it to another agent, or release it if no target is provided. It is categorised as a Write tool in the Agent Bus MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Agent Bus MCP server in PolicyLayer and add a rule for handoff_task: 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 Agent Bus. Nothing to install.
handoff_task is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the handoff_task 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 handoff_task. 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.
handoff_task is provided by the Agent Bus MCP server (mustaphasteph/agent-bus). 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 →