Skills
skills/orchestra-research/ai-research-skills/ray-data/Gen Agent Trust Hub

ray-data

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [Unverifiable Dependencies] (MEDIUM): The skill specifies ray[data] and pandas as dependencies. These are hosted by organizations (ray-project, pandas-dev) not included in the pre-approved trusted list defined in the analysis guidelines.
  • [Indirect Prompt Injection] (LOW): The skill describes processes for ingesting data from cloud storage (S3, GS, Parquet) which could contain malicious instructions.
  • Ingestion points: ray.data.read_parquet, read_csv, and read_json in SKILL.md.
  • Boundary markers: Absent from provided transformation and loading examples.
  • Capability inventory: Support for arbitrary distributed Python execution across compute clusters via the map_batches and map methods.
  • Sanitization: No data validation, schema enforcement, or content sanitization logic is shown for data entering the pipeline.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 06:05 PM