phase-6-finalize

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill accepts a session_id input (a session-scoped token) and explicitly instructs forwarding it verbatim in command arguments and agent prompt bodies (e.g., "--session-id {session_id}" / prompt fields), which forces the LLM to handle and output a secret value directly.