winston-geo-onpage: per-URL on-page GEO action lists with screenshots and JSON-LD.

What it does.

Feed it a list of URLs. The Skill scrapes each page, analyzes how the content chunks for AI citation, scores the existing schema, drafts the Schema.org JSON-LD recommendation, captures a screenshot, and writes the per-URL action list. Output is a recommendation deck for stakeholders, a schema sheet for the dev team, an implementation checklist, and a content priority tracker.

How we use it.

It runs after the GEO Technical Audit, when the strategist has decided which pages to optimize. Triggered URL-by-URL or in batch. The strategist reviews the recommendations, edits the priority call, and routes the deliverables to the client's content and dev teams. The implementation checklist becomes the actual work order.

Why we built it.

Per-page GEO work was eating two hours of senior time per URL: scrape, read, chunk-analyze, draft schema, screenshot, write the recommendation, file the priority. Across an engagement that's 30 to 100 URLs of senior time. The judgment lives in the priority call and the strategic which-pages-matter decision. The Skill handles the rest.

The stack inside it.

• Firecrawl for content scrape and screenshot capture
• Anthropic API and Claude Skills for chunking analysis and schema drafting
• Schema.org reference for JSON-LD generation per page type
• openpyxl for the schema sheet and priority tracker
• python-pptx for the per-URL recommendation deck
• A reusable JSON-LD template library covering Article, Product, FAQ, HowTo, LocalBusiness

What this proves.

Skills compound. Every job that was a half-day or a multi-day stretch of senior time becomes a Skill the team triggers from chat. The output ships faster. The strategist's calendar opens up for the work that actually needs human judgment: positioning, narrative, the prioritization call on what the data is really saying. That is the Winston operating model in one Skill.