AI agents call verify_bundle to retrieve information from Verify without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves and validates data from a signed artifact (audit bundle) without creating, modifying, deleting, or executing any code. It is a pure inspection/query operation. The offline nature and use of embedded keys confirm no external state is modified. Severity is low because verification failures have no operational impact beyond reporting validation status.
From the tool's definition Tool performs offline verification of audit bundles using 'embedded verification keys' — a read-only cryptographic check operation with no side effects. The verb 'verify' indicates validation/inspection, not modification or execution of arbitrary code.
Attacks that exploit this kind of access
Verify a ScopeBlind audit bundle offline using the embedded verification keys. It is categorised as a Read tool in the Verify MCP Server, which means it retrieves data without modifying state.
Register the Verify MCP server in PolicyLayer and add a rule for verify_bundle: 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 Verify. Nothing to install.
verify_bundle 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 verify_bundle 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 verify_bundle. 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.
verify_bundle is provided by the Verify MCP server (scopeblind/verify-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 →