tool_compare_runs
AI agents call tool_compare_runs to retrieve information from Inspect Logs without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool operates within a log analysis and evaluation inspection system. Given the name 'compare_runs' in the context of comparing evaluation test runs, this almost certainly retrieves and compares historical log data without modifying it. However, confidence is moderate (0.6) because the description is empty, limiting certainty about the exact implementation and scope.
From the tool's definition Tool name 'tool_compare_runs' and server context show this belongs to the inspect-logs-mcp server which provides capabilities to 'explore and analyze UK Government BEIS inspect_ai evaluation logs'.
Attacks that exploit this kind of access
tool_compare_runs. It is categorised as a Read tool in the Inspect Logs MCP Server, which means it retrieves data without modifying state.
Register the Inspect Logs MCP server in PolicyLayer and add a rule for tool_compare_runs: 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 Inspect Logs. Nothing to install.
tool_compare_runs 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 tool_compare_runs 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 tool_compare_runs. 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.
tool_compare_runs is provided by the Inspect Logs MCP server (pranshusrivastava/inspect-logs-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 →