seo-fix-plan

Pass

Audited by Gen Agent Trust Hub on Jul 3, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill interacts with UnifAPI endpoints (including browser/* and seo/*) to fetch live search and website data. These are vendor-provided services necessary for the skill's core functionality.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from external websites (HTML/Markdown) to identify SEO issues. While this presents a theoretical attack surface, the skill's design is restricted to generating a textual report ("eyes not hands"), which prevents automated exploitation or unauthorized system changes.
  • [COMMAND_EXECUTION]: The provided skill files do not contain any shell commands, subprocess calls, or system-level execution patterns.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 3, 2026, 04:06 PM
Security Audit — agent-trust-hub — seo-fix-plan