The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute several Python scripts using shell commands that point to hardcoded absolute Windows file paths (e.g., python C:\Users\renat\skills\007\scripts\surface_mapper.py). This grants the skill the ability to run code on the host system, which is a significant security risk if the environment or the scripts are compromised.
[PROMPT_INJECTION]: As a security auditing tool, the skill ingests untrusted code and data from external projects, creating a vector for indirect prompt injection. Malicious instructions within the analyzed files could attempt to override the agent's logic or exfiltrate information.
Ingestion points: Target source code, configuration files, and project directories analyzed by the skill.
Boundary markers: The skill does not define clear delimiters or use specific instructions to prevent the agent from following directions embedded in the analyzed data.
Capability inventory: The skill has access to shell execution, file system reading, and various CLI tools including claude-code and cursor.
Sanitization: There is no evidence of sanitization, filtering, or sandboxing of the content being audited.

007