sumi-orchestrator
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill serves as an entry point for processing diverse user requests related to design tasks.
- Ingestion points: User UX/UI requests provided at runtime (SKILL.md).
- Boundary markers: None explicitly defined in the orchestration logic to delimit external data.
- Capability inventory: The orchestrator triggers multi-stage pipelines that involve file system read/write operations to the
.sumi/directory for state management via thedesign-memoryskill. - Sanitization: Includes a mandatory ethical safety gate (ux-ethics-content-strategy) to identify and filter manipulative content or dark patterns within the generated design recommendations.
Audit Metadata