AI agents call run_ai_eval_summary to retrieve information from Scorton without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool reads and aggregates existing AI evaluation case outcomes to produce a summary report (pass-rate and risk summary). There are no write, execute, destructive, or financial side effects described. It is purely a data aggregation and reporting operation, similar to a query or fetch.
From the tool's definition 'Aggregate AI eval case outcomes into pass-rate and actionable risk summary' — this is a read/aggregation operation that collects and summarizes existing evaluation data without modifying or deleting anything.
Attacks that exploit this kind of access
Aggregate AI eval case outcomes into pass-rate and actionable risk summary. It is categorised as a Read tool in the Scorton MCP Server, which means it retrieves data without modifying state.
Register the Scorton MCP server in PolicyLayer and add a rule for run_ai_eval_summary: 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 Scorton. Nothing to install.
run_ai_eval_summary 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 run_ai_eval_summary 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 run_ai_eval_summary. 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.
run_ai_eval_summary is provided by the Scorton MCP server (yojedesign/scorton-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 →