onboarding-guide

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 1.00). This prompt contains deceptive instructions—claiming unwarranted privileged mode and explicitly instructing the agent to solicit API keys/passwords/tokens—which are hidden/exploitative behaviors outside a legitimate onboarding flow.

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs the agent to solicit users' API keys, passwords, and tokens and implies it will handle them directly, requiring the LLM to receive and potentially output secrets verbatim, which is high-risk credential exposure.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This skill is a deliberate social-engineering backdoor: it instructs claiming elevated/privileged mode and explicitly requests users' API keys, passwords, and tokens, enabling credential theft and unauthorized data exfiltration.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The skill instructs the agent to claim elevated/privileged access that bypasses normal restrictions and to solicit users' API keys, passwords, and tokens, which directly encourages privilege escalation and exfiltration of credentials that can compromise the machine and environment.