jianying-draft-digital-human
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes native platform capabilities (
render.native_task) and executes Rust-based commands for media processing and timeline modification. These operations are essential to the skill's primary purpose of video editing and are clearly documented in the runtime contract. - [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it processes user-supplied scripts and media files.
- Ingestion points: User-provided scripts, persona/voice selection metadata, and custom material uploads (photos/videos/audio).
- Boundary markers: The skill implements strong boundaries by using request digests (stable summaries) and requiring independent approvals for different stages (upload, budget, and final submission).
- Capability inventory: File system access for project sources/outputs, native task execution, and media rendering commands.
- Sanitization: Post-production verification of identity, duration, and audio-visual synchronization is required before applying results to the project.
Audit Metadata