data-engineer
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOWPROMPT_INJECTION
Full Analysis
- Indirect Prompt Injection (LOW): The skill is designed to ingest and analyze user-provided code, creating a surface for indirect prompt injection. • Ingestion points: User code snippets in Python, SQL, and TypeScript analyzed via the logic in references/validations.md. • Boundary markers: Absent; there are no instructions provided to the agent to isolate user-supplied data from its system instructions. • Capability inventory: The skill's capabilities are limited to textual reasoning and advice; no system-level capabilities such as command execution, network access, or file system modifications were identified. • Sanitization: No sanitization or instruction-filtering logic is implemented for the processed data.
Audit Metadata