Hi, I'm internet-dot with HOL. Your README says "manage_api_keys - Create, list, get, update, and delete org-wide or project-scoped API keys". We'd like Guard to surface that access change for review before an agent runs the MCP tool. Would you consider contributing a kernel-mcp-server extension to HOL Guard?

HOL Guard checks AI agents' commands before execution. You could require approval for manage_api_keys while keeping read-only operations available to the agent. You define and test that boundary. The extension builder can draft a rule and tests from an exported MCP tool list. You'd decide which invocations need review and send the PR to hashgraph-online/hol-guard. No changes to this repo or its CI are needed. The extension stays off unless a user enables it.

For the PR: start with an exported MCP tool list, review the generated rule, add cases for both approval-required and read-only invocations, and run the handoff check. The guide includes a source example.