Hey Product Hunt! agentools.sh is live there today. Tell me what to add →
Sponsor

Data skills

13 Data skills, most installed first. I read every file in each one to see what it can access, and show the security audits from skills.sh. See all skills.

Best for Solo founders who want their coding agent to handle databases, spreadsheets and data

This page lists the Agent Skills I filed under data: database rules, spreadsheets, analytics tracking, Amazon and Google search data, social post research and even photo geolocation. They are sorted by total installs on skills.sh. I picked each one and read it file by file to rate its access (Low, Medium or High), and every skill page shows the security audits from skills.sh.

If you run on Postgres, start with Postgres Best Practices. It is the most installed skill here, it has Low access, and it shows wrong and right SQL for schema, indexes, RLS and query tuning. For spreadsheets, Excel Spreadsheets (XLSX) insists on live formulas and uses a recalc script to report broken cells. To study a competitor, try outlier-post-finder, which flags the posts that beat an account's own baseline.

1Postgres Best Practicessupabase/agent-skillsPostgres schema, RLS and query rulesLow3/3 pass463K2Supabasesupabase/agent-skillsBuild on Supabase without RLS mistakesHigh1 warn347K3Excel Spreadsheets (XLSX)anthropics/skillsSpreadsheets with real, error-free formulasMedium1 warn189K4last30daysmvanhorn/last30days-skillResearches what people actually said about any topic in the last 30 days across Reddit, X, YouTube, HN and moreHigh1 fail57.2K5EAS Update Insightsexpo/skillsCrash rates and adoption for published updatesHigh3/3 pass37.3K6EAS Observeexpo/skillsApp metrics and error reporting with expo-observeHigh3/3 pass19.5K7geo-sleuthOldcircle/geo-sleuthFind where a photo was taken, with evidence for every claimHigh3/3 pass3.3K8Scientific Agent SkillsK-Dense-AI/scientific-agent-skills177 research skills for science agentsHigh3/3 pass2.5K9outlier-post-finderScrapeCreators/social-media-research-skillsFind the posts that beat a creator's own baseline, and whyHigh1 warn1.3K10PostHog Product AnalyticsPostHog/ai-pluginAdd PostHog event tracking to your appHigh2 warn59511Pangolinfo AI SERPpangolinfo/pangolinfo-skillsLets your agent search Google and pull AI Overviews with their citations, plus keyword trends, as clean JSONHigh1 warn112Pangolinfo Amazon Scraperpangolinfo/pangolinfo-skillsTeaches your agent to pull live Amazon products, keyword results, bestseller lists and reviews as clean JSONHigh1 warn113NotFair marketing skillsnowork-studio/notfair-plugin48 SEO, GEO and paid ads skills that answer with evidence, not vibesHighNo audits—

Questions

What is the best Agent Skill for Postgres in Claude Code?

I'd start with Postgres Best Practices from Supabase. It has Low access, is MIT licensed, and covers schema design, indexes, RLS policies, connection pooling, locking and query tuning for any Postgres, not only Supabase. Install it with npx skills add supabase/agent-skills --skill supabase-postgres-best-practices.

What is the difference between the Supabase skill and Supabase Postgres Best Practices?

Postgres Best Practices is a set of SQL rules for any Postgres database and has Low access. Supabase covers every Supabase product, with a security checklist for RLS, views, SECURITY DEFINER functions, keys and storage, plus CLI and MCP workflows for migrations and debugging. Its access is High.

How do I install the Excel XLSX skill in Claude Code?

Run /plugin marketplace add anthropics/skills, then /plugin install document-skills@anthropic-agent-skills. Excel Spreadsheets (XLSX) works with Claude Code, Claude.ai and the Claude API, and its license is proprietary (source-available).

Can an AI agent pull Amazon product data and Google AI Overviews as JSON?

Yes. Pangolinfo Amazon Scraper pulls products, keyword results, bestseller lists and reviews, and Pangolinfo AI SERP pulls Google results, AI Overviews with their citations and Google Trends comparisons. Both need the Pangolinfo Amazon Data MCP connected and a Pangolinfo API key.