tools-only TODO — claude-seo
Completed in v1.2.0
- Fix YAML frontmatter parsing — Removed HTML comments before
---in 8 files (from @kylewhirl fork) - SSRF prevention in Python scripts — Private IP blocking in fetch_page.py and analyze_visual.py (from @artyomsv #7)
- Install hardening — venv-based pip, no
--break-system-packages(from @JawandS #2) - Windows install fixes —
python -m pip,py -3fallback, requirements.txt persistence (from @kfrancis #5, PR #6) - requirements.txt persistence — Copied to skill dir after install (from @edustef #1)
- Path traversal prevention — Output path sanitization in capture_screenshot.py, file validation in parse_html.py
Deferred from Community Feedback
Reduce Bash scope on agents (Priority: Medium, from @artyomsv #7) Evaluate which agents truly need Bash access. Consider replacing with WebFetch where possible.
Docker-based script execution (Priority: Low, from @artyomsv #7) Sandbox Python scripts in Docker for users who want extra isolation.
Opencode compatibility (Priority: Low, from @Ehtz #4) Adapt skill architecture for Opencode. @kylewhirl already ported to OpenAI Codex.
Subagent timeout/compact handling (Priority: Medium, from @JawandS #3) Primary agent sometimes terminates before subagents finish. Consider encouraging subagents to run /compact and adding explicit wait logic.
Native Chrome tools vs Playwright (Priority: Medium, from @artyomsv #7, @btafoya PR #8) Claude Code has native browser automation. Evaluate replacing Playwright with built-in tools to eliminate the ~200MB Chromium dependency.
Deferred from February 2026 Research Report
Fake freshness detection (Priority: Medium) Compare visible dates (
datePublished,dateModified) against actual content modification signals. Flag pages with updated dates but unchanged body content.Mobile content parity check (Priority: Medium) Compare mobile vs desktop meta tags, structured data presence, and content completeness. Flag discrepancies that could affect mobile-first indexing.
Discover optimization checks (Priority: Low-Medium) Clickbait title detection, content depth scoring, local relevance signals, sensationalism flags.
Brand mention analysis Python implementation (Priority: Low) Currently documented in
seo-geo/SKILL.mdbut no programmatic scoring.
Last updated: February 19, 2026