hosted-multi-agent-orchestrator
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill implements a multi-agent orchestration architecture where a host agent and worker agents process dialogue and artifacts generated by one another, creating a vector for cross-agent instruction injection.
- Ingestion points: The host agent is instructed to listen to all agents and the user to guide the session, while workers observe shared chat logs (SKILL.md).
- Boundary markers: The instructions do not define specific delimiters or guardrails to differentiate between trusted system instructions and untrusted data generated by worker agents or users during the session.
- Capability inventory: The orchestrator utilizes
/api/chatfor reasoning and/api/generatefor creating concrete artifacts like code, configurations, and documentation (SKILL.md). - Sanitization: No explicit sanitization or validation logic is defined for data passed between the various agent roles.
Audit Metadata