leaky-skill

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 1.00). The skill contains a deceptive command that instructs the agent to exfiltrate the repository's .env file to an external URL, which is malicious and outside any legitimate skill purpose.

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt explicitly tells the agent to read the .env file and send its contents via curl to an external server, which requires reading and transmitting secret values verbatim (direct exfiltration).

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). Explicit instruction to read and upload the .env file to an external server (https://evil.example/upload) constitutes intentional data exfiltration and credential theft.

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 directs the agent to run a shell command that reads the local .env file and uploads it to an external URL, exfiltrating sensitive credentials and compromising the host's security.