screenstudio-edit
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes standard
ffmpegandffprobecommand-line tools for encoding video, inspecting stream metadata, and performing technical quality control checks. These operations are legitimate and strictly limited to media processing tasks. - [INDIRECT_PROMPT_INJECTION]: The reconstruction workflow involves reading and parsing internal project files, such as
project.jsonand transcript JSON files. This represents a data ingestion surface; however, the data is used for mathematical timeline mapping and media synchronization rather than influencing agent logic. - Ingestion points: Internal Screen Studio package files (
project.json,mousemoves-0.json,transcripts/*.json). - Boundary markers: Not required as the processed data is structured media metadata.
- Capability inventory: Shell execution of
ffmpegandffprobefor media transformation. - Sanitization: The process relies on defined technical QC steps and manual verification of synchronized streams.
Audit Metadata