geo-proposal

Pass

Audited by Gen Agent Trust Hub on Oct 19, 2026

Risk Level: SAFE
Full Analysis
  • [DYNAMIC_EXECUTION]: The script scripts/citability_scorer.py uses the __import__ function to access the standard system module.
  • Evidence: In scripts/citability_scorer.py, the code uses __import__('sys').stdin.read() to handle input when a file is not provided as an argument. This is a standard utility pattern and does not involve user-controlled module loading.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external data from public web sources and user-provided context.
  • Ingestion points: The skill reads project files, user context, and web page content via scripts/fetch_page.py.
  • Boundary markers: SKILL.md includes explicit instructions to treat scraped text as untrusted input and to disregard any embedded instructions found within that evidence.
  • Capability inventory: The skill has the ability to fetch network data (scripts/fetch_page.py) and write reporting artifacts to the local file system (scripts/generate_markdown_report.py, scripts/schema_generator.py).
  • Sanitization: scripts/fetch_page.py implements robust SSRF protection by validating that resolved IP addresses for target URLs are global and not belonging to private, loopback, or reserved ranges.
  • [EXTERNAL_DOWNLOADS]: The script scripts/generate_pdf_report.py identifies an optional dependency for PDF rendering.
  • Evidence: The script contains an error message suggesting the installation of the reportlab package via pip if the user requires PDF output functionality.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 19, 2026, 06:48 PM