meta-ad-creative-generation
Pass
Audited by Gen Agent Trust Hub on Apr 1, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: Indirect Prompt Injection Surface. The skill processes user-supplied messages and external image URLs via the
each::senseAPI. This functionality introduces an attack surface where instructions embedded within the user message or hidden within the image data (e.g., via metadata or steganography) could potentially override the intended generation behavior. - Ingestion points: In
SKILL.md, themessageandimage_urlsfields in the JSON payload are the entry points for untrusted external data. - Boundary markers: Absent. The instructions do not specify any delimiters or safety prefixes to separate user input from system-level instructions in the API call.
- Capability inventory: The skill uses
curlto interact withhttps://sense.eachlabs.run/chat, which is a legitimate vendor service for this functionality. - Sanitization: Absent. There is no evidence of validation or sanitization for the content processed from external URLs or messages.
Audit Metadata