vaex

Pass

Audited by Gen Agent Trust Hub on Jun 27, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill consists of technical reference material and code snippets for the Vaex library. All provided functionality aligns with legitimate data science and big data processing tasks.
  • [PROMPT_INJECTION]: The skill identifies an indirect prompt injection surface due to its core functionality of ingesting large datasets from untrusted sources (e.g., CSV, Parquet, S3 buckets). * Ingestion points: Data loading methods such as vaex.open() and vaex.from_csv() documented in core_dataframes.md and io_operations.md. * Boundary markers: No specific markers or delimiters are mentioned to isolate data content from instructions. * Capability inventory: Operations include file system writing (export_hdf5), network access for cloud storage (s3fs, gcsfs), and data visualization. * Sanitization: No explicit data validation or sanitization steps are defined for processing external data.
  • [SAFE]: Dynamic execution is restricted to performance-enhancing features like Numba JIT compilation for custom calculations, which is a standard and safe implementation within the Vaex ecosystem.
  • [SAFE]: No attempts to bypass safety filters, exfiltrate credentials, or establish persistence were found.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 27, 2026, 11:43 AM
Security Audit — agent-trust-hub — vaex