langsmith-dataset

Warn

Audited by Snyk on Mar 4, 2026

Risk Level: MEDIUM
Full Analysis

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

  • Third-party content exposure detected (high risk: 0.70). The skill explicitly loads user-provided trace JSONL files (load_traces_from_dir / load_traces_from_file in generate_datasets.py/ts) and calls the LangSmith API to list/read dataset examples (query_datasets.py/ts uses Client.read_dataset and list_examples), thereby ingesting user-generated, third-party dataset content that the agent parses and displays as part of its runtime workflow.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 4, 2026, 12:32 AM