storyboard-cheatcode
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided scene concepts, art styles, and negative constraints, which are then interpolated into complex prompt templates for image and video generation tools.
- Ingestion points: User inputs for CONCEPT, PANEL ART STYLE, and HARD NEGATIVES defined in Step 1 of SKILL.md.
- Boundary markers: The skill does not employ explicit delimiters or 'ignore embedded instructions' markers when wrapping user input into the final tool prompts.
- Capability inventory: The skill utilizes the
higgsfield_generate_imageandhiggsfield_generate_videotools to perform generations on the Higgsfield platform. - Sanitization: There is no evidence of input validation or escaping before interpolation into the generation prompts.
- [EXTERNAL_DOWNLOADS]: The skill provides instructions for the user to add an external Model Context Protocol (MCP) server from the official Higgsfield domain (https://mcp.higgsfield.ai/mcp). This is a standard configuration requirement for the skill to interact with its intended generation service.
Audit Metadata