AI agents use wb_plan to create or update resources in Workbench — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Workbench environment.
The tool creates and manages task assignments derived from page-PRDs, which is a reversible write operation (tasks can be cancelled/modified). The auto-cancel on page deletion touches task state but is a side effect of the write/dispatch action rather than a standalone destructive operation. Idempotency further suggests controlled write semantics.
From the tool's definition 派发模块整组任务(幂等,删页自动 cancel) — dispatches/creates a group of tasks based on registered page-prd, and auto-cancels tasks when pages are deleted
Attacks that exploit this kind of access
按已登记 page-prd 派发模块整组任务(幂等,删页自动 cancel). It is categorised as a Write tool in the Workbench MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Workbench MCP server in PolicyLayer and add a rule for wb_plan: 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 Workbench. Nothing to install.
wb_plan 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 wb_plan 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 wb_plan. 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.
wb_plan is provided by the Workbench MCP server (nvrenshiren/workbench). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
wb_plan is one line of Workbench's registry record.
The record carries the whole server: verified identity, auth posture, risk grade, every tool classified, recommended policy — re-checked continuously.
Teams ship this data inside their own products. See what a licence covers →