The Agent Skills Directory

COMMAND_EXECUTION (SAFE): The skill uses grep and find to detect the project environment. These are read-only operations targeting local metadata (package.json and directory structure) to provide context-aware recommendations. They do not execute external code or modify the system.
DATA_EXPOSURE (SAFE): No evidence of credential leakage or unauthorized data exfiltration was found. The code templates correctly demonstrate best practices by using environment variables (process.env) for secrets rather than hardcoding them.
REMOTE_CODE_EXECUTION (SAFE): The skill does not perform any network operations, download external scripts, or execute untrusted code.
PROMPT_INJECTION (SAFE): No patterns of prompt injection, role-play overrides, or instructions to bypass safety filters were detected in the skill body or metadata.

auth-expert