persona-hello-world
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data retrieved from the Persona API, such as inquiry status and verification results.
- Ingestion points: Data enters the agent context through
requests.getandrequests.postcalls to the Persona API inSKILL.md. - Boundary markers: The instructions lack delimiters or explicit warnings to the agent to disregard potential instructions embedded in the API responses.
- Capability inventory: The skill allows access to
Bash,Read,Write, andEdittools, creating a potential attack surface if the API returns malicious content. - Sanitization: The Python code snippets print raw data from the API response to the console without validation or filtering.
Audit Metadata