export
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill directs the agent to interact with the local file system by resolving and writing to standard user directories (
~/Downloadsor%USERPROFILE%\Downloads). This is performed to save exported media files, manage naming collisions, and monitor browser download artifacts (e.g.,.crdownloadfiles). - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources, such as video project timelines, source assets, and NLE XML files. It possesses capabilities to write to the file system and perform network downloads based on this data, which represents a potential injection surface.
- Ingestion points: Project assets, timeline state data, and imported XML files (SKILL.md).
- Boundary markers: No explicit delimiters or instructions to ignore embedded commands within the project data are provided.
- Capability inventory: File system write access to the user's Downloads folder and network download capabilities via the
request_asset_downloadandsubmit_exporttools. - Sanitization: The instructions do not specify any sanitization or validation routines for ingested project metadata or XML content.
Audit Metadata