AI agents call agloop_get_state to retrieve information from Agloop without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves and queries internal agent framework state (phase, tasks, iteration, compaction context) with no side effects. It is a pure read operation that provides introspection into an active AgLoop instance. The low severity reflects that unauthorized state access alone does not directly enable damaging operations, though detailed internal state could inform subsequent attacks.
From the tool's definition Tool description explicitly states 'Read the full AgLoop state' and returns state data without modification. No write, execute, delete, or financial operations described.
Attacks that exploit this kind of access
Read the full AgLoop state including phase, tasks, iteration, and compaction context. Returns null if no active loop. It is categorised as a Read tool in the Agloop MCP Server, which means it retrieves data without modifying state.
Register the Agloop MCP server in PolicyLayer and add a rule for agloop_get_state: 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 Agloop. Nothing to install.
agloop_get_state is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the agloop_get_state 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 agloop_get_state. 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.
agloop_get_state is provided by the Agloop MCP server (zebbern/agloop-mcp). 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 →