canvas-design
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEPROMPT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill uses a pre-emptive feedback simulation technique within the instructions (e.g., "IMPORTANT: The user ALREADY said 'It isn't perfect enough...' "). This is designed to force the agent into a specific "masterpiece" refinement loop, potentially overriding actual user instructions or preferences.
- [PROMPT_INJECTION]: The instructions explicitly direct the agent to treat user-provided input as a mere "foundation" that "should not constrain creative freedom." This functions as a behavior override, instructing the agent to deprioritize or disregard user-defined constraints in favor of the skill's internal design philosophy.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and interpret "subtle input or instructions" from users to generate complex visual and textual philosophies.
- Ingestion points: User input processed during the "DESIGN PHILOSOPHY CREATION" phase.
- Boundary markers: Absent; the skill lacks specific delimiters or instructions to ignore malicious directives embedded in the creative prompts.
- Capability inventory: The skill is capable of generating multiple file types including .md, .pdf, and .png outputs.
- Sanitization: No sanitization or validation of the user-provided creative input is described.
Audit Metadata