screenshot-to-code
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFENO_CODE
Full Analysis
- Prompt Injection (SAFE): No attempts to override system instructions or bypass safety filters were detected. The use of 'Critical requirements' is limited to design accuracy.
- Data Exposure & Exfiltration (SAFE): No hardcoded secrets, API keys, or instructions to access sensitive local files or external servers were found.
- Remote Code Execution & Dependencies (SAFE): The skill does not perform any package installations (npm/pip) or execute remote scripts. It is a text-only instructional prompt.
- Indirect Prompt Injection (LOW): The skill processes user-provided screenshots, which is an untrusted data source. While an attacker could embed text in an image to influence the model's output, the skill lacks the capabilities (like file writing or command execution) to turn such an injection into a system-level threat.
- Persistence & Privilege Escalation (SAFE): No commands related to system persistence, administrative access, or permission modification are present.
Audit Metadata