csv-parse

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill provides functionality to ingest and process CSV data from external sources, creating an attack surface where untrusted data could influence agent behavior.
  • Ingestion points: Several functions in SKILL.md (parseCsvFile, parseCsvString, and validateCsv) are designed to read data from external files or input strings.
  • Boundary markers: The implementation does not include explicit boundary markers or instructions to the agent to ignore potentially embedded instructions within the CSV fields.
  • Capability inventory: The skill utilizes Node.js fs streams for reading from and writing to the local file system.
  • Sanitization: The provided examples do not perform sanitization or content filtering on the parsed CSV data before it is returned as objects for further use.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 10:40 AM
Security Audit — agent-trust-hub — csv-parse