The Agent Skills Directory

[PROMPT_INJECTION] (SAFE): No instructions attempting to override the agent's system prompt or bypass safety filters were detected. The markdown content consists solely of documentation and report templates.
[DATA_EXFILTRATION] (SAFE): The code performs all calculations locally. No network-based exfiltration patterns (e.g., HTTP requests via requests or urllib) or access to sensitive credential files were found.
[REMOTE_CODE_EXECUTION] (SAFE): The skill does not use dangerous functions like eval(), exec(), or os.system() on user-supplied input. Logic is contained within statically defined modules.
[OBFUSCATION] (SAFE): All source code and metadata are provided in cleartext. There is no evidence of Base64 encoding, zero-width characters, or homoglyph attacks used to hide malicious intent.
[COMMAND_EXECUTION] (SAFE): Command line interaction is managed via argparse, which is a standard and secure library for CLI utilities. No unsafe subprocess spawning was observed.

title-expert