wolf-howl

Fail

Audited by Snyk on Feb 27, 2026

Risk Level: HIGH
Full Analysis

HIGH W007: Insecure credential handling detected in skill instructions.

  • Insecure credential handling detected (high risk: 0.90). The setup explicitly shows passing a wallet value and chat ID as CLI arguments (python3 ... --wallet {WALLET} --chat-id {CHAT_ID}) and states the agent "already knows" these values, which implies the model would need to insert secret/credential values verbatim into generated cron/command output — an exfiltration risk.
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 27, 2026, 09:43 PM