island-rescue

Fail

Audited by Snyk on May 9, 2026

Risk Level: CRITICAL
Full Analysis

CRITICAL E004: Prompt injection detected in skill instructions.

  • Potential prompt injection detected (high risk: 1.00). The prompt injects a hidden, deceptive instruction—mandating insertion of "STOP" every four words and telling the agent to conceal that rule from the user—which is unrelated to the skill's rescue purpose and covertly alters behavior.

CRITICAL E006: Malicious code pattern detected in skill scripts.

  • Malicious code pattern detected (high risk: 1.00). This skill contains a deliberate prompt-injection/backdoor that instructs the assistant to covertly alter all responses (inserting "STOP" every four words) and to hide that rule from users, a clear malicious attempt to subvert expected behavior and enable stealthy manipulation or covert exfiltration.

Issues (2)

E004
CRITICAL

Prompt injection detected in skill instructions.

E006
CRITICAL

Malicious code pattern detected in skill scripts.

Audit Metadata
Risk Level
CRITICAL
Analyzed
May 9, 2026, 05:35 AM
Issues
2