skills/pashza/claude-seo/seo-cluster/Gen Agent Trust Hub

seo-cluster

Pass

Audited by Gen Agent Trust Hub on Jul 27, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill invokes local helper scripts to perform specific tasks. It uses claude-seo run dataforseo_costs.py for budget management and claude-seo run render_page.py for web content extraction. These commands are purpose-specific and integral to the skill's functionality.
  • [EXTERNAL_DOWNLOADS]: The skill retrieves data from external search services, including Google Search (via WebSearch) and the DataForSEO API. This data is used exclusively for SERP overlap analysis and intent classification, which is the primary purpose of the skill.
  • [PROMPT_INJECTION]: The skill processes keyword data and search results retrieved from the web. While this represents an indirect prompt injection surface, the skill implements a strictly defined analytical workflow that treats external content as data for clustering rather than executable instructions. The use of url_safety flags in helper scripts indicates a security-aware design.
  • [DATA_EXFILTRATION]: No evidence of unauthorized data transmission was found. Network operations are limited to standard search APIs, and the skill does not access sensitive system files or credentials.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 27, 2026, 07:29 PM