NYC

multi-source-data-merger

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [Indirect Prompt Injection] (LOW): The skill is designed to process data from external, heterogeneous sources which constitutes an ingestion surface for untrusted data.
  • Ingestion points: Heterogeneous source files including JSON, CSV, Parquet, and XML mentioned in SKILL.md.
  • Boundary markers: None specified for the external data content.
  • Capability inventory: Python script execution, file read/write operations (e.g., creating output/merged_data.json).
  • Sanitization: Not explicitly addressed; the skill focuses on data normalization and type conversion for functional purposes rather than security.
  • [Dynamic Execution] (LOW): The workflow involves generating and executing Python scripts to perform data merging.
  • Evidence: The skill guides the agent to write a script and verify it using 'python -m py_compile script.py' and subsequent execution.
  • Context: This is a standard approach for data engineering tasks and uses logic derived from the skill's own templates.
  • [Unverifiable Dependencies] (SAFE): The skill recommends installing standard, widely-trusted Python libraries.
  • Evidence: Explicitly mentions 'pandas', 'pyarrow', 'openpyxl', and 'lxml' via 'uv pip install'.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:04 PM