Business data analysis skill for AI coding agents. Turns raw data into actionable business intelligence — from CSV exploration to professional PDF reports. Works with Claude Code, Gemini CLI, Codex CLI, OpenCode, and any agent that reads ~/.claude/skills/.
/askprisma runs a structured 5-phase analysis workflow:
- Data Discovery — profiles all data files or SQL tables (row counts, types, nulls, cardinality, statistics)
- Analysis Planning — presents a numbered plan tailored to your data and question; waits for approval
- Iterative Execution — writes and runs Python code task by task, fixing errors automatically
- Business Synthesis — translates findings into plain language with specific numbers and "so what" implications
- PDF Report Generation — optionally generates a professional PDF in executive summary, comprehensive, or slide format
npx askprisma-skillDetects which AI tools are present and copies the skill into each one's standard directory:
| Agent | Install Path |
|---|---|
| Claude Code / cowork | ~/.claude/skills/askprisma/ |
| Gemini CLI | ~/.gemini/extensions/askprisma/ |
| Codex CLI | ~/.codex/skills/askprisma/ |
| OpenCode | ~/.config/opencode/skills/askprisma/ |
npx skills add whiteboardmonk/askprisma-skillOr install for a specific agent only:
npx skills add whiteboardmonk/askprisma-skill --agent claude
npx skills add whiteboardmonk/askprisma-skill --agent cursorPreview what will be installed without installing:
npx skills add whiteboardmonk/askprisma-skill --listgit clone https://github.com/whiteboardmonk/askprisma-skill
ln -s "$PWD/askprisma-skill/skills/askprisma" ~/.claude/skills/askprisma/askprisma What are the main trends in sales_data.csv?
/askprisma Which product categories are driving churn this quarter?
/askprisma Segment our customers and generate an executive summary PDF
After analysis completes, the agent will ask if you want a PDF. Choose from:
- Executive Summary — 1-2 pages for leadership
- Comprehensive Report — full analysis with methodology
- Slide Presentation — visual PDF in 16:9 format
All artifacts are saved to ./askprisma-outputs/ in your working directory:
*.png— charts and visualizations*.csv— intermediate data tablesreport_[style]_[YYMMDDHHMM].pdf— PDF reports
pip install -r requirements.txtOr individually:
pip install pandas numpy matplotlib seaborn scipy scikit-learn statsmodels openpyxl reportlab Pillowaskprisma-skill/
├── bin/
│ └── install.js # npx cross-platform installer
├── .claude-plugin/
│ └── plugin.json # Claude plugin manifest
├── skills/
│ └── askprisma/
│ ├── SKILL.md # Skill instructions (frontmatter + content)
│ ├── references/
│ │ ├── coding-patterns.md # Python coding conventions
│ │ ├── business-translation.md # Technical-to-business translation guide
│ │ └── report-styles.md # PDF report JSON schemas
│ └── scripts/
│ └── generate_report.py # PDF generator (standalone CLI)
├── gemini-extension.json # Gemini CLI extension manifest
├── package.json # npm package (agents.skills field for npx skills add)
├── LICENSE
├── README.md
├── CHANGELOG.md
└── requirements.txt
python skills/askprisma/scripts/generate_report.py \
--input /tmp/test_report.json \
--output /tmp/test_output.pdf \
--charts-dir /tmpMIT — see LICENSE.