higgsfield-character-design
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process potentially large amounts of external, user-provided data, specifically 'story bibles' and character sheets stored in the
workspace/input/directory. - Ingestion points: The agent is instructed to load and read 'canon reference' documents and character descriptions from the local workspace to maintain story consistency.
- Capability inventory: The skill acts as a pre-production layer that 'routes' its outputs to other skills like
higgsfield-promptandhiggsfield-soul. These downstream skills likely perform more active generations based on the inputs gathered here. - Boundary markers: The instructions do not explicitly mandate the use of delimiters (e.g., XML tags or triple backticks) when the agent processes or interpolates data from the story bible into prompt generation steps.
- Sanitization: There are no explicit instructions for the agent to sanitize or validate the content of the external bible files before using them to influence subsequent agent actions.
Audit Metadata