fei-fei-li
Pass
Audited by Gen Agent Trust Hub on Apr 25, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: No attempts to override agent behavior, bypass safety guardrails, or extract system prompts were detected. The instructions focus entirely on adopting a specific reasoning style (human-centered AI and spatial intelligence).
- [DATA_EXFILTRATION]: No sensitive file access or credential harvesting patterns were found. All network-related content consists of legitimate citations and references to well-known domains like Stanford University, LinkedIn, and major media outlets.
- [OBFUSCATION]: No malicious encoding or hidden content was discovered. Standard Unicode characters for punctuation in JSON data are present but used conventionally for typographic purposes.
- [REMOTE_CODE_EXECUTION]: The skill does not contain any instructions to download, install, or execute external packages or scripts. It is composed of static text and structured data files.
- [COMMAND_EXECUTION]: No evidence of shell command execution, privilege escalation (e.g., sudo), or persistence mechanisms was found in any of the files.
- [INDIRECT_PROMPT_INJECTION]: The skill does not possess an attack surface for indirect injection; it provides a fixed corpus of knowledge and does not ingest untrusted runtime data for interpolation into sensitive tools.
Audit Metadata