autonomous-agents

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • NO_CODE (SAFE): The skill consists entirely of markdown documentation and instructional prompts. It does not include any scripts (Python, JS, Bash) or executable logic.
  • EXTERNAL_DOWNLOADS (SAFE): No external URLs, package managers (npm/pip), or remote scripts are referenced or executed.
  • DATA_EXFILTRATION (SAFE): No commands for file access or network transmission are present. The content actually recommends security guardrails and least privilege.
  • PROMPT_INJECTION (SAFE): The instructions define a persona (Agent Architect) for educational purposes and do not attempt to bypass safety filters, extract system prompts, or override agent constraints.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:45 PM