paddle-distributed

Pass

Audited by Gen Agent Trust Hub on Mar 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is entirely composed of documentation and reference guides for PaddlePaddle developers. No security threats or malicious patterns were detected.
  • [PROMPT_INJECTION]: The instructions and documentation do not contain any patterns aimed at overriding agent behavior or bypassing safety guidelines.
  • [DATA_EXFILTRATION]: No hardcoded credentials, sensitive file paths, or unauthorized network operations were identified.
  • [EXTERNAL_DOWNLOADS]: The skill does not include any external package installations (pip, npm) or remote script executions (curl|bash).
  • [COMMAND_EXECUTION]: There are no instructions that would lead to unauthorized shell command execution or system-level persistence.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 13, 2026, 07:55 AM