designing-agents

Pass

Audited by Gen Agent Trust Hub on Apr 5, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns or security vulnerabilities were detected in the skill files. The content is strictly educational and advisory in nature.
  • [PROMPT_INJECTION]: No attempts to bypass safety filters, override system instructions, or extract system prompts were found.
  • [DATA_EXFILTRATION]: No network operations, credential exposures, or access to sensitive local files were identified.
  • [REMOTE_CODE_EXECUTION]: The skill does not download or execute external scripts or packages.
  • [COMMAND_EXECUTION]: While the skill discusses giving agents access to shell and filesystem as a design pattern for other systems, it does not include any direct or hidden command execution instructions for its own implementation.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 5, 2026, 02:56 AM
Security Audit — agent-trust-hub — designing-agents