microsoft-agent-framework

Pass

Audited by Gen Agent Trust Hub on May 4, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is purely informational, providing high-level steps for role separation, workflow control, and governance in agent systems.
  • [EXTERNAL_DOWNLOADS]: The skill includes a reference to the official Microsoft GitHub repository for the agent framework, which is a well-known and trusted source.
  • [PROMPT_INJECTION]: No instructions were found that attempt to bypass safety guidelines, override system prompts, or extract internal configuration.
  • [COMMAND_EXECUTION]: Although the skill is configured with access to tools like Bash, it does not include any automated scripts or command-line instructions that could be executed maliciously.
Audit Metadata
Risk Level
SAFE
Analyzed
May 4, 2026, 12:15 PM
Security Audit — agent-trust-hub — microsoft-agent-framework