voice-ai

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (critical risk: 1.00). The skill instructs the user to execute an unvetted remote npx installer script from a third-party GitHub repository without review or pinning, which runs unverified code.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs runtime installation via an npx command pulling an unvetted personal GitHub repository, presenting a weak/unknown external dependency.