alterlab-genai-image-to-video
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions direct the agent to actively ingest data from external sources and local files, creating a potential attack surface for indirect prompt injection.
- Ingestion points: The
WorkflowandAgentic Protocolsections require the agent to "Search the web for current Soul ID capabilities, video model updates" and to "Read existing project files for context — storyboards, shot lists, image asset inventories". - Boundary markers: There are no instructions for the agent to use boundary markers (like delimiters or XML tags) or specific prompts to ignore potential instructions embedded within the data it reads.
- Capability inventory: The skill involves capabilities for web searching and extensive local file writing (e.g.,
{project}-shot-plan.md,{project}-sequence-plan.md), which increases the potential impact if the agent is successfully manipulated by external content. - Sanitization: The instructions do not include any steps for sanitizing or validating the content retrieved from external web searches or project files before it is processed or used to generate outputs.
Audit Metadata