The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes the ffprobe utility via the subprocess.run Python method. The implementation correctly passes arguments as a list rather than a shell string, which is a best practice that prevents shell injection vulnerabilities even when processing user-provided file paths.
[INDIRECT_PROMPT_INJECTION]: While the skill processes external data (MP4 files and JSON plans), the risk of indirect prompt injection is mitigated by the 'Iron Rule' in the instructions which forbids the agent from reading raw data files directly. Instead, the agent only consumes a summarized, validated report produced by the local script.
[DATA_EXPOSURE]: The skill reads project-specific files (plan.json and media files) required for its primary task. It does not access sensitive system locations, environment variables, or credentials.

highlight-verifier