scenario-kling
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill manages video generation tasks that ingest user-provided text prompts and media assets, creating an interface for potential indirect prompt injection attacks.\n
- Ingestion points: Untrusted content is processed through prompts and asset uploads in the
model_runandupload_assettools (SKILL.md).\n - Boundary markers: The skill employs markers like
@Element1,<<<image_1>>>, and<<<video_1>>>to reference external assets within the prompt structure.\n - Capability inventory: The skill allows the agent to search for models, retrieve execution schemas, upload files, and trigger complex video/audio generation jobs.\n
- Sanitization: No explicit sanitization or verification of user-provided content is defined within the skill's instructions.
Audit Metadata