data-engineering-storage-remote-access-libraries-fsspec

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS] (SAFE): Installation of established Python libraries (fsspec, s3fs, gcsfs, adlfs) from PyPI.
  • [DATA_EXFILTRATION] (LOW): The skill demonstrates reading from and writing to remote storage and external URLs (e.g., https://example.com). This is the intended purpose but constitutes a network access surface.
  • [PROMPT_INJECTION] (LOW): The skill provides a surface for indirect prompt injection. 1. Ingestion points: Remote data files loaded via fsspec and pandas from HTTPS/S3. 2. Boundary markers: No delimiters or instructions provided to ignore embedded content in data. 3. Capability inventory: Includes file system manipulation (mkdir, rm) and network reads. 4. Sanitization: No sanitization of ingested content before processing.
  • [CREDENTIALS_UNSAFE] (SAFE): Documentation uses the standard 'AKIA...' placeholder for AWS keys; no real secrets are exposed.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:32 PM