SelfSubjectAccessReview: can this SA perform <verb> on <resource>?
AI agents call check_permission to retrieve information from Gke Cred Audit without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
SelfSubjectAccessReview is a read-only authorization check that returns yes/no answers about permissions without triggering any side effects, state changes, or code execution. Even though it relates to security-sensitive RBAC information, the tool itself only retrieves and reports existing permission status.
From the tool's definition Tool performs a SelfSubjectAccessReview check, which is a read-only Kubernetes API operation that queries whether a service account has permission to perform an action.
Attacks that exploit this kind of access
SelfSubjectAccessReview: can this SA perform <verb> on <resource>?. It is categorised as a Read tool in the Gke Cred Audit MCP Server, which means it retrieves data without modifying state.
Register the Gke Cred Audit MCP server in PolicyLayer and add a rule for check_permission: 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 Gke Cred Audit. Nothing to install.
check_permission 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 check_permission 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 check_permission. 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.
check_permission is provided by the Gke Cred Audit MCP server (rrupesh/mcp-test). 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 →