Start a run-until-done goal loop. Records the completion condition; the agent should then
AI agents use set_goal to create or update resources in Goal Engine — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Goal Engine environment.
This tool creates or modifies persistent goal state (reversibly via clear_goal). It does not execute arbitrary code, delete data, or move money. The severity is medium because a malicious agent could set goals that cause the loop to behave unexpectedly, but the effects are reversible and the tool itself does not directly cause side effects—only records intent that the agent then acts upon.
From the tool's definition Tool name 'set_goal' and description 'Start a run-until-done goal loop. Records the completion condition' indicate the tool creates and persists new goal state that modifies the server's tracking system.
Attacks that exploit this kind of access
Start a run-until-done goal loop. Records the completion condition; the agent should then. It is categorised as a Write tool in the Goal Engine MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Goal Engine MCP server in PolicyLayer and add a rule for set_goal: 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 Goal Engine. Nothing to install.
set_goal 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 set_goal 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 set_goal. 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.
set_goal is provided by the Goal Engine MCP server (melihzafer/mcp-goal). 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 →