Skills
Navigate
skills/reason-machines/data-skills/harvard-art-museum-data-engineering/Gen Agent Trust Hub

harvard-art-museum-data-engineering

Pass

Audited by Gen Agent Trust Hub on Oct 1, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructions include cloning a source code repository from GitHub (https://github.com/Manali0711/Harvard-Artifacts-Collection-Data-Engineering-Analytics-App.git) and installing several third-party Python dependencies.
  • Evidence: The 'Installation' section specifies git clone and pip install -r requirements.txt for setup.
  • [INDIRECT_PROMPT_INJECTION]: The skill fetches artifact metadata (titles, descriptions, culture, etc.) from the Harvard Art Museums API and processes it into a database and Streamlit dashboard.
  • Ingestion points: Data is retrieved via the requests library from api.harvardartmuseums.org in the fetch_artifacts function.
  • Boundary markers: The skill does not implement specific boundary markers or instructions for the agent to ignore embedded instructions within the API data.
  • Capability inventory: The skill has the capability to write to SQL databases (mysql.connector), read files (dotenv), and render content in a web dashboard (streamlit).
  • Sanitization: While the code performs basic string slicing ([:500]) to fit database schema constraints, it does not sanitize content for potential malicious prompt fragments embedded in artifact metadata.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 1, 2026, 01:35 PM
Security Audit — agent-trust-hub — harvard-art-museum-data-engineering