ag2-network-workflow
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is an educational resource providing boilerplate code and architectural patterns for the AG2 network workflow system.
- [INDIRECT_PROMPT_INJECTION]: The skill describes an orchestration system that processes messages and packets between multiple agents. While this creates a surface where one agent could potentially influence another through malicious instructions in message content, this is inherent to the multi-agent workflow use case. The documentation focuses on routing logic rather than input sanitization at the orchestration layer. Ingestion points involve turn-taking envelopes and capability inventory includes tool execution via @agent.tool as described in SKILL.md.
- [DYNAMIC_EXECUTION]: The framework allows users to register custom subclasses for transition targets and conditions. This mechanism facilitates framework extensibility by allowing the instantiation of local code at runtime, following standard object-oriented design patterns for plugin systems.
Audit Metadata