harvard-artifacts-collection-analytics-pipeline

Warn

Audited by Gen Agent Trust Hub on May 24, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the agent to clone a repository from https://github.com/Manali0711/Harvard-Artifacts-Collection-Data-Engineering-Analytics-App.git. This source is a personal account and is not a verified organization or well-known service.
  • [COMMAND_EXECUTION]: Executes shell commands to set up the project environment, specifically git clone, pip install -r requirements.txt, and streamlit run app.py.
  • [PROMPT_INJECTION]: The skill ingests artifact metadata, media, and color data from the Harvard Art Museums API (api.harvardartmuseums.org), which is an untrusted external source. This creates a surface for indirect prompt injection if the ingested content is used to influence subsequent agent actions or logic.
  • Ingestion points: Fetches data records in the fetch_artifacts_from_api function within SKILL.md.
  • Boundary markers: No delimiters or instructions are used to distinguish the ingested API data from the agent's core instructions.
  • Capability inventory: The skill has database write capabilities via mysql-connector-python and network access via requests.
  • Sanitization: While the code uses SQL parameterization for data values, it does not validate or sanitize the textual content of the API response for potential malicious instructions before processing.
Audit Metadata
Risk Level
MEDIUM
Analyzed
May 24, 2026, 06:49 AM
Security Audit — agent-trust-hub — harvard-artifacts-collection-analytics-pipeline