The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes Python scripts (check_provenance.py, dump_metadata.py, extract_frames.py) that wrap external forensic tools like c2patool, exiftool, and ffmpeg. These scripts use subprocess.run with arguments passed as a list and without enabling shell execution, which effectively prevents command injection. This is the intended functionality of the skill and is implemented securely.
[EXTERNAL_DOWNLOADS]: The skill references several well-known and reputable external services for media verification, including Google DeepMind, the Content Authenticity Initiative, and ElevenLabs. It also provides links to official Israeli government resources for reporting cyber incidents and harm to minors. These references are informative and do not involve untrusted or risky downloads.
[SAFE]: No malicious patterns, such as prompt injection, data exfiltration, or obfuscation, were detected. The skill's instructions focus strictly on its stated purpose of media verification and provide clear, honest guidance on the limitations of the tools and methods used.

israeli-media-authenticity-verifier