byted-data-quality-review-loop
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process arbitrary deliverables, including code and documents, which are then passed to a sub-agent via the
sessions_spawntool for review. This creates a surface where instructions embedded in the user-provided data could potentially influence the sub-agent's review process. \n - Ingestion point: 'Current deliverable' (SKILL.md). \n
- Boundary markers: None explicitly defined to separate data from instructions. \n
- Capability inventory: Uses
sessions_spawnfor agent orchestration. \n - Sanitization: No specific filtering or sanitization of input data is mentioned. \n- [SAFE]: The skill defines a legitimate and structured multi-round quality review process. It is authored by the vendor (ByteDance) and implements a standard business logic for quality gating. No unauthorized file access, network exfiltration, or persistence mechanisms were found.
Audit Metadata