rowlab-labeling-at-scale

Pass

Audited by Gen Agent Trust Hub on Nov 7, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for processing large external datasets, which introduces a potential surface for indirect prompt injection if the datasets contain malicious instructions.
  • Ingestion points: Large user-provided datasets (SKILL.md).
  • Boundary markers: Absent; the instructions do not specify delimiters or warnings to ignore instructions embedded in the data.
  • Capability inventory: The agent is instructed to perform SQL queries and make API calls to rowlab endpoints to process the dataset.
  • Sanitization: No specific sanitization or validation of the data rows is prescribed.
  • [NO_CODE]: The skill is comprised entirely of instructional Markdown content and does not include any scripts, shell commands, or binary files.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 7, 2026, 11:53 AM