write-pipeline
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines patterns for ingesting and processing data from external sources, creating a potential surface for indirect prompt injection. * Ingestion points: Scripts ingest data from external APIs (e.g., fetch_charges) and user-provided S3 files via data_upload. * Boundary markers: The instructions do not recommend using delimiters or warnings to ignore instructions embedded in the ingested data. * Capability inventory: The pipelines can execute SQL queries and write to workspace object storage (S3/DuckLake). * Sanitization: While the skill promotes parameterized queries to prevent SQL injection, it does not address the sanitization of content that might contain instructions for an LLM.
Audit Metadata