Skills
Navigate
skills/gingiris-1031/gingiris-skills/gr-seo-patrol/Gen Agent Trust Hub

gr-seo-patrol

Pass

Audited by Gen Agent Trust Hub on Oct 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from external websites which are potentially attacker-controlled ingestion points.
  • Ingestion points: scripts/check-page.py, scripts/check-schema.py, scripts/check-social.py, scripts/check-site.py, and scripts/daily-report.py fetch and parse external web content (H1 tags, meta descriptions, JSON-LD, robots.txt, etc.).
  • Boundary markers: SKILL.md includes a 'HARD RULE' anti-hallucination guardrail requiring the agent to only report on fields defined in the scripts' JSON output envelopes.
  • Capability inventory: The agent can modify files in a GitHub repository via scripts/canonical-fix.py and scripts/rescue-post.py, and it can perform network operations.
  • Sanitization: No specific filtering or sanitization of the extracted string content is performed before it is presented to the agent.
  • [COMMAND_EXECUTION]: The skill provides shell command templates for running Python scripts involving user-supplied URLs (e.g., python3 scripts/check-page.py URL).
  • [EXTERNAL_DOWNLOADS]: The scripts utilize requests and urllib.request to fetch data from user-defined URLs, the DataForSEO API (api.dataforseo.com), and the GitHub API (api.github.com).
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 15, 2026, 08:12 AM