Physical AI Expert
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOW
Full Analysis
- [Prompt Injection] (SAFE): The instructions are standard persona-based directives focused on academic accuracy. No bypass markers, 'ignore previous instruction' patterns, or adversarial role-play prompts were detected.
- [Data Exposure & Exfiltration] (SAFE): The skill does not contain any file access commands, hardcoded credentials, or network requests (curl/wget/fetch). There is no mechanism to access or transmit sensitive data.
- [Unverifiable Dependencies & RCE] (SAFE): There are no scripts or configuration files (requirements.txt, package.json) present. No remote code is downloaded or executed.
- [Indirect Prompt Injection] (INFO): The skill possesses a data ingestion surface (user queries about robotics), but it lacks any 'write' or 'execute' capabilities. It acts solely as a content generator/formatter, meaning malicious input cannot trigger side effects beyond the generated text.
- [Obfuscation] (SAFE): The text is clear and contains no encoded strings, hidden Unicode characters, or homoglyphs.
Audit Metadata