kerberos-delegation

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt includes plaintext example passwords, NTLM hashes, and API-like keys and explicitly instructs embedding credentials/hashes in commands (e.g., getTGT.py, Rubeus, printerbug.py, addcomputer.py), requiring the LLM to handle or output secret values verbatim — a high exfiltration risk.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This document is an explicit, actionable offensive playbook for stealing Kerberos tickets and credentials, coercing authentication, creating attacker machine accounts, modifying AD objects (RBCD), and using stolen tickets for remote impersonation, privilege escalation and lateral movement — all high-risk malicious behaviors.

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

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The skill explicitly instructs creating and modifying AD/computer accounts (RBCD, addcomputer.py), extracting credentials from LSASS (Mimikatz/Rubeus) which requires local admin, and even suggests running sudo ntpdate (root), all of which modify or require privileged access to the host/domain state.