AI agents call axle_extract_theorems to retrieve information from Axle without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The name 'extract' combined with 'theorems' indicates data retrieval from Lean proof objects, consistent with Read category. No destructive, financial, or code execution semantics are evident. The empty description reduces confidence, but the naming pattern aligns with analysis/retrieval tools in the Lean proof engineering context. Confidence lowered due to lack of explicit description.
From the tool's definition Tool name 'extract_theorems' suggests retrieval/querying of theorem data. Description is empty, limiting direct evidence. Similar sibling tools (axle_check, axle_normalize, axle_rename) appear to be read or analysis operations on Lean proofs.
Attacks that exploit this kind of access
axle_extract_theorems. It is categorised as a Read tool in the Axle MCP Server, which means it retrieves data without modifying state.
Register the Axle MCP server in PolicyLayer and add a rule for axle_extract_theorems: 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 Axle. Nothing to install.
axle_extract_theorems 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 axle_extract_theorems 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 axle_extract_theorems. 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.
axle_extract_theorems is provided by the Axle MCP server (vilin97/axle-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 →