The Agent Skills Directory

[SAFE]: The skill functions as an educational resource, providing best practices for secure file management. It includes correct implementations for magic byte validation, filename sanitization, and secure presigned URL handling.- [COMMAND_EXECUTION]: No unauthorized or suspicious command execution patterns were found. The code snippets provided are intended for application-level logic and use safe, standard libraries.- [DATA_EXFILTRATION]: The skill does not contain any instructions or code that would result in the exfiltration of sensitive data. It correctly advises on using short-lived credentials and proper cache control for cloud storage access.- [INDIRECT_PROMPT_INJECTION]: While the skill discusses handling untrusted user input (filenames and file content), it does so to instruct the agent on how to secure the application against such inputs rather than creating a vulnerability surface itself.

file-uploads