cloud-storage-optimization
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill implements functionality to ingest and process external data files (CSV and Parquet), which serves as an attack surface where adversarial instructions could be embedded to influence the agent's actions.
- Ingestion points: Data is loaded from local file paths and S3 sources in
references/data-compression-and-partitioning-strategy.mdusingpandas.read_csvand inreferences/data-lake-partitioning-strategy.mdusingpyarrow.parquet.read_table. - Boundary markers: The code lacks explicit delimiters or instructions to the agent to treat ingested data as non-executable content, increasing the risk of the agent following instructions found within the data.
- Capability inventory: The skill has significant capabilities including AWS S3 bucket configuration (lifecycle policies, intelligent tiering), multi-cloud infrastructure management via Terraform, and local filesystem write access in the
/tmpdirectory. - Sanitization: No sanitization, schema validation, or content filtering is performed on the data before it is processed or used to generate cloud storage keys and paths.
Audit Metadata