nimrodfisher/data-analytics-skills

What it solves

Traditional AI assistants need heavy upfront configuration—schemas, metric definitions, business rules—before they can help with data work. This project removes that barrier. It gives you a ready-made set of 31 portable skills that Claude or Codex can pull in on demand, covering the entire analyst workflow: checking data quality, digging into patterns, documenting findings, building dashboards, and explaining results to stakeholders. No setup, no company-specific rules, and it works for any industry.

How it works

Each skill is a folder with a SKILL.md file (plus optional scripts, references, and assets) that follows the open Agent Skills standard. You clone the repo and copy the skills you want into your agent's skills directory. When you describe a task naturally—like "why did our activation rate drop 12%?"—the agent automatically picks the right skill and runs it. The skill follows an on-demand context pattern: it asks only for the minimum information needed, executes a structured step-by-step workflow, flags any assumptions it's making, and produces a consistent, shareable output. If you can't provide everything, it degrades gracefully—it states the assumptions and proceeds.

Who it's for

Data analysts, data scientists, and anyone who uses Claude Code or Codex (or compatible agents like Cursor or Windsurf) for hands-on analytic work. It's built for people who want to go from raw data to clear insights without spending days on configuration. It's also useful for teams that need consistent, repeatable analytics processes across projects or companies.

Highlights

• 31 skills across 6 categories: data quality & validation, documentation & knowledge, analysis & investigation, storytelling & visualization, stakeholder communication, and workflow optimization.
• Zero setup required: works out-of-the-box, no schemas or business rules to predefine.
• On-demand activation: describe your task and the agent picks the right skill automatically.
• Portable across any organization: not tied to a specific company or industry.
• Customizable: add a references/ folder to include company-specific schema, metric definitions, or business rules.
• Chainable: combine 4-5 skills end-to-end for full projects.
• Structured workflow: each skill asks targeted questions, executes a clear process, and surfaces assumptions.
• Graceful degradation: if you can't provide full context, the skill states its assumptions and proceeds.