signup-flow-cro
Pass
Audited by Gen Agent Trust Hub on Mar 7, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: No security threats detected. The skill is purely instructional and provides marketing advice based on user input.
- [PROMPT_INJECTION]: The instructions follow standard agent persona and task-specific guidelines without any attempts to bypass safety filters or override behavior. Regarding indirect injection: Ingestion points include '.agents/product-marketing-context.md'; boundary markers and sanitization are absent; however, the capability inventory confirms no dangerous tools (network, file-write, or command execution) are present, ensuring safety.
- [DATA_EXFILTRATION]: No sensitive file paths or exfiltration patterns were found; access to marketing context files is standard for this agent type.
- [NO_CODE]: The skill consists of documentation and evaluation test cases only, with no executable scripts or system-level commands.
Audit Metadata