skills/lev-os/agents/geo/Gen Agent Trust Hub

geo

Pass

Audited by Gen Agent Trust Hub on Mar 21, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it retrieves and processes content from untrusted external websites provided by the user.\n
  • Ingestion points: scripts/fetch_page.py, scripts/citability_scorer.py, and scripts/llmstxt_generator.py use the requests library to fetch HTML content from external URLs.\n
  • Boundary markers: The analysis logic lacks explicit delimiters or instructions to the agent to disregard potential commands embedded in the fetched HTML content.\n
  • Capability inventory: The skill has the ability to write files (GEO-AUDIT-REPORT.md, GEO-REPORT.pdf), execute local scripts via Bash, and fetch further web content.\n
  • Sanitization: Content is parsed for structure but not sanitized to remove or escape potential natural language instructions.\n- [COMMAND_EXECUTION]: The orchestration logic in SKILL.md and the PDF generation process involve executing local Python scripts within the agent's environment.\n
  • Evidence: SKILL.md references the execution of scripts/generate_pdf_report.py to compile audit data.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 21, 2026, 06:49 PM
Security Audit — agent-trust-hub — geo