seo-geo

Pass

Audited by Gen Agent Trust Hub on May 12, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill performs its stated purpose of SEO analysis without any malicious intent. The code is well-structured and uses standard, secure libraries.
  • [EXTERNAL_DOWNLOADS]: The skill requires standard Python dependencies listed in requirements.txt. These are common libraries such as requests and beautifulsoup4 and include version pinning for security.
  • [COMMAND_EXECUTION]: The skill utilizes local Python scripts (scripts/fetch_page.py and scripts/parse_html.py) to process data. This is a legitimate use of tool execution for the skill's functionality.
  • [PROMPT_INJECTION]: The skill ingests and analyzes content from external URLs provided by the user. While this creates a surface for indirect prompt injection where malicious instructions could be embedded in the analyzed page's text, this is a standard risk for web-parsing tools and is managed by the model's core safety layers.
  • Ingestion points: External URLs fetched via scripts/fetch_page.py as described in SKILL.md.
  • Boundary markers: None explicitly defined in the instructions for content analysis.
  • Capability inventory: Execution of local Python scripts fetch_page.py and parse_html.py.
  • Sanitization: Content is parsed using BeautifulSoup to extract structural and text elements before analysis.
Audit Metadata
Risk Level
SAFE
Analyzed
May 12, 2026, 03:09 AM
Security Audit — agent-trust-hub — seo-geo