eve-orchestration
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The orchestration pattern involves parent agents generating child job descriptions that may incorporate untrusted external data, which could contain hidden instructions targeting child agents.\n
- Ingestion points: Parent agents read and synthesize summaries and work products from child jobs as described in the orchestration flow in SKILL.md.\n
- Boundary markers: The child job description template uses structural headers (Scope, Context, Deliverable) but lacks explicit instructions for child agents to ignore instructions embedded within the ingested context.\n
- Capability inventory: The skill leverages the
eveCLI for job management and execution within a bash environment, which provides significant capability to any successfully injected instructions.\n - Sanitization: The instructions do not prescribe sanitization or escaping of data before it is passed between jobs or interpolated into job descriptions.
Audit Metadata