cueframe-scene-shot
Pass
Audited by Gen Agent Trust Hub on Nov 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill operates through platform-defined tools (
get_scene_contract,apply_composition) rather than arbitrary shell commands, ensuring actions stay within the intended application context. - [SAFE]: It includes defensive instructions that explicitly prohibit the use of raw HTML, external URLs, or unauthorized shader source code in scene data fields.
- [SAFE]: The workflow requires the agent to fetch a canonical schema and use specific media IDs registered through a controlled workflow, preventing the ingestion of untrusted external content into sensitive processing paths.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from internal platform resources (
cueframe://scene-shot) and tool outputs (get_scene_contract). While this is an ingestion surface, it is controlled by the platform's refusal rules and schema validation. Ingestion points: cueframe://scene-shot resource and get_scene_contract tool output. Boundary markers: strict JSON schema. Capability inventory: apply_composition, scene.add, scene.update. Sanitization: platform-level budget and schema enforcement.
Audit Metadata