ideogram
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from
brand-profile.mdand user-supplied text to construct complex image generation prompts. - Ingestion points: The skill reads brand guidelines, palettes, and typography rules from
brand-profile.md(SKILL.md) and accepts user strings for in-image text. - Boundary markers: Instructions specify using JSON structures and quotes to delimit text elements (references/json-prompting-and-layout.md), which offers partial protection against accidental instruction execution.
- Capability inventory: The skill interfaces with the Ideogram API and manages a workflow that uploads generated assets to an external media platform (tools/integrations/ideogram.md).
- Sanitization: There is no mention of sanitizing or validating the contents of the brand profile file or user-provided strings before interpolation.
Audit Metadata