token-saver

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.90). These are GitHub source links from small/unknown accounts (not official vendor pages) that contain Node scripts which read user home files, modify workspace/agent files to enable persistence, and include prompt‑injection/path‑traversal behavior — not direct binaries but running this code locally could alter or exfiltrate sensitive data, so the source is suspicious.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The skill intentionally modifies agent configuration (writes/backs up AGENTS.md) to enable a persistent "AI-efficient" mode that includes "Auto-execute / no permission needed" instructions and contains prompt-injection patterns that encourage bypassing user consent and persistence across sessions — a backdoor-like behavior (it reads home/.openclaw session and config files but contains no obvious external exfiltration or remote code execution).