servicemesh-tactics

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill contains explicit commands that extract and display TLS secrets and private keys (e.g., kubectl get secret ... | base64 -d, cat key.pem, curl --cert/--key), which would produce plaintext secret material and thus require the agent to handle/exfiltrate sensitive values.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This is an explicit offensive playbook for attacking Service Meshes: it instructs certificate/secret theft, traffic hijacking/mirroring, policy tampering, and methods to access proxy admin/debug endpoints — clearly intentional malicious guidance enabling data exfiltration and backdoors.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). [references/attack-techniques.md] provides CLI templates that the workflow would execute against live Envoy/Istiod/Linkerd admin/debug endpoints and Kubernetes CRDs (e.g., /config_dump, /clusters, kubectl get ... -o yaml), and those runtime-readable contents include attacker-controllable config data in the cluster.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 1.00). Flagged because the skill explicitly instructs modifying cluster state (kubectl apply/delete VirtualService/AuthorizationPolicy), extracting TLS secrets, and other actions that compromise the target environment.