ecommerce-amazon-opportunity-report-by-keyword

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and display external Markdown reports generated by an AI-powered API, creating a surface for potential instructions to be embedded in the report content.
  • Ingestion points: The skill accepts a user-provided keyword and retrieves a stdout field containing Markdown text from the NexScope API (scripts/amazon_opportunity_report.py).
  • Boundary markers: The skill lacks explicit structural delimiters for the external content but includes a directive: "This tool's output cannot be fed into data query tools for further processing."
  • Capability inventory: The skill's Python script performs network POST requests and writes JSON data to the local file system under the nexscope/ directory.
  • Sanitization: There is no evidence of sanitization or filtering of the Markdown content before it is displayed by the agent.
  • [COMMAND_EXECUTION]: The skill relies on the execution of a Python script to handle its core logic, including API communication and data management.
  • The scripts/amazon_opportunity_report.py script is executed with user-controlled parameters passed as a JSON string.
  • [EXTERNAL_DOWNLOADS]: The skill performs network operations to communicate with the vendor's API infrastructure.
  • The Python script uses urllib.request.urlopen to send search parameters to the endpoint defined by the NEXSCOPE_PROXY_BASE environment variable.
  • These network operations are necessary for the skill's primary function of retrieving market opportunity reports.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 04:12 AM
Security Audit — agent-trust-hub — ecommerce-amazon-opportunity-report-by-keyword