create-end-card-ffmpeg
Pass
Audited by Gen Agent Trust Hub on Aug 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/create_end_card.pyexecutesffmpegandffprobeusing thesubprocess.runmethod. The use of argument lists prevents shell injection, but the skill still exercises significant control over system processes. - [INDIRECT_PROMPT_INJECTION]: The skill exhibits a potential surface for indirect injection through untrusted data interpolation.
- Ingestion points: The skill accepts
text_linesandlogo_optionsas JSON inputs which are directly used to build video filter chains. - Boundary markers: No delimiters or safety instructions are used to separate untrusted text from the logic of the filter graph.
- Capability inventory: The skill possesses capabilities for file system writes and binary execution (
ffmpeg). - Sanitization: Minimal sanitization is performed, specifically escaping single quotes and colons for the
drawtextfilter. - [EXTERNAL_DOWNLOADS]: The skill documentation and implementation reference external dependencies, including the Python packages
cairosvgandPillow, as well as system-level libraries for image processing.
Audit Metadata