higgsfield-brand-story
Pass
Audited by Gen Agent Trust Hub on Jun 18, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill processes user-provided brand intent and narrative descriptions to construct prompts for video generation, which creates a surface for indirect prompt injection. However, the skill implements a human-in-the-loop safety pattern.
- Ingestion points: User input for brand story type, emotional register, and target platform in
SKILL.md('Gather intent' step). - Boundary markers: Explicit instruction in
SKILL.md('Present for review' step) to show the final prompt and parameters for user approval before any tool call. - Capability inventory:
mcp__higgsfield__generate_videoandmcp__higgsfield__generate_imagetools inSKILL.mdfrontmatter. - Sanitization: The skill relies on explicit user opt-in and prompt verification before executing generation tools.
Audit Metadata