animated-scene-still
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data in the form of image references and user-supplied text to construct structured JSON image prompts.
- Ingestion points: User-provided image handles (for character, environment, and prop identity) and specific text strings for signs or packaging are processed in
SKILL.mdand defined in thereferences/json-blueprint.mdtemplate. - Boundary markers: The skill uses a structured JSON blueprint (
references/json-blueprint.md) to categorize inputs, which provides some structural isolation, but it lacks explicit instructions to ignore potentially malicious commands embedded within the reference data. - Capability inventory: The skill's functionality is limited to text generation. No file-system access, network operations, or command execution capabilities were identified in the analyzed files.
- Sanitization: There are no documented mechanisms for sanitizing or validating the content of the reference images or the exact text strings provided by the user.
- [NO_CODE]: The skill consists exclusively of Markdown instructions and a JSON template. It does not include any scripts, binaries, or automated tasks that could execute logic on the host system.
Audit Metadata