design-prompt-builder
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill creates an attack surface by collecting untrusted user input and interpolating it into prompt templates intended for use in other AI agents.
- Ingestion points: User-provided context such as 'Product Description', 'Target Persona', and 'Design Hypotheses' is captured via interactions defined in
SKILL.mdand synthesized into the final output prompt. - Boundary markers: The skill uses markdown headers and separator lines (e.g.,
---) to structure the prompt, but it does not provide specific instructions to the receiving agent to treat the[CONTEXT]section as untrusted or to ignore any malicious instructions embedded within that data. - Capability inventory: The skill itself has no execution capabilities (it does not perform file system writes, network requests, or shell command execution); its sole function is generating text for the user to review and copy-paste.
- Sanitization: There is no evidence of input validation, filtering, or escaping of the user-provided strings before they are interpolated into the resulting prompt template.
Audit Metadata