Report Generator
Transform raw data into polished, executive-ready reports with AI-powered analysis, Mermaid charts, and optional multi-agent pipelines.
Overview
The Report Generator turns your data files, Notion databases, and Google Drive documents into professional Markdown reports -- complete with executive summaries, trend analysis, and embedded visualizations. Whether you have a CSV of quarterly sales figures or a folder of project metrics, Founder OS reads the data, runs statistical analysis, writes narrative prose, and delivers a report you can hand to stakeholders.
The namespace supports two modes. The default single-agent mode produces a 3-5 page report in one pass -- ideal for quick internal updates. The --team flag activates a full 5-agent pipeline (Research, Analysis, Writing, Formatting, QA) that produces deeper analysis with richer visualizations and a quality review pass. Team mode also records the finished report in your Founder OS HQ - Reports database in Notion.
Data sources are auto-detected: CSV, JSON, plain text, Markdown, log files, Notion databases, and Google Drive documents (via the gws CLI). You can point the command at a single file or an entire directory -- the system processes each source, merges results where appropriate, and highlights the most important findings.
Required Tools
| Tool | Required | Purpose |
|---|---|---|
| Filesystem | Yes | Read data source files and write report output |
| Notion CLI | No | Record reports in the HQ Reports database (team mode) |
| gws CLI | No | Extract data from Google Drive documents |
Commands
/founder-os:report:generate
What it does -- Generates a polished Markdown report from one or more data sources. In default mode, it reads your data, performs statistical analysis, and writes a concise report with an executive summary, key findings, Mermaid charts, and recommendations. With --team, it runs the full 5-agent pipeline for deeper analysis and a QA review pass.
Usage:
/founder-os:report:generate [--data=PATH] [--output=PATH] [--team]
Example scenario:
You have a CSV of Q4 sales data and need to brief your advisory board tomorrow. Run
/founder-os:report:generate --data=sales-q4.csvand you get a report with revenue trends, top-performing segments, anomaly flags, and prioritized recommendations -- ready to share in under a minute.
What you get back:
A Markdown file in your output directory containing an executive summary, key findings with supporting metrics, 1-2 Mermaid chart visualizations, and actionable recommendations. The output summary shows the file path, data sources processed, section count, chart count, and a preview of the executive summary.
Flags:
--data=PATH-- Path to a data file or directory. If omitted, you are prompted for a source.--output=PATH-- Output directory (default:./report-output/)--team-- Activate the full 5-agent pipeline with Notion recording
/founder-os:report:from-template
What it does -- Generates a report using a predefined template that controls section layout and formatting. Three templates are available: executive-summary for quick stakeholder updates, full-business-report for quarterly deep dives, and project-status-report for RAG-status milestone tracking. Template placeholders are filled with your data, and chart markers are replaced with Mermaid visualizations.
Usage:
/founder-os:report:from-template [--template=NAME] [--data=PATH] [--output=PATH] [--team]
Example scenario:
Your team runs weekly sprint retrospectives and you want a consistent format every time. Run
/founder-os:report:from-template --template=project-status-report --data=sprint-metrics.jsonand the output follows your standard layout -- RAG status, milestone progress, action items -- with this week's data filled in automatically.
What you get back:
A Markdown report following the chosen template structure exactly, with data-derived content filling each section. If no template is specified, you see a list of available templates with descriptions to choose from.
Flags:
--template=NAME-- Template name:executive-summary,full-business-report, orproject-status-report. If omitted, lists available templates.--data=PATH-- Path to data source. If omitted, you are prompted.--output=PATH-- Output directory (default:./report-output/)--team-- Activate the full 5-agent pipeline with Notion recording
The 5-Agent Pipeline (Team Mode)
When you use --team, the report passes through five specialized agents in sequence:
| Order | Agent | Role |
|---|---|---|
| 1 | Research Agent | Gathers and extracts data from all specified sources |
| 2 | Analysis Agent | Computes statistics, detects trends, identifies outliers and correlations |
| 3 | Writing Agent | Generates report prose with executive summary and narrative data interpretation |
| 4 | Formatting Agent | Adds Mermaid charts, formats tables, polishes the final output |
| 5 | QA Agent | Reviews for accuracy, consistency, and data integrity (recommendation-only) |
Each agent passes structured JSON output to the next. The pipeline produces a richer report than single-agent mode and includes a QA review with flags for anything that warrants a second look.
Supported Data Sources
| Source Type | How It Works |
|---|---|
.csv files |
Auto-detects delimiters, headers, and column types |
.json files |
Supports arrays and nested objects with automatic flattening |
.txt / .md / .log files |
Pattern-matches tables, key-value pairs, and log entries |
| Notion databases | Queries via Notion CLI with dynamic database discovery |
| Google Drive | Exports Sheets as CSV, Docs as text via gws CLI |
| Directories | Scans for all supported file types and processes each |
What Data Formats Can the Report Generator Process?
The Report Generator auto-detects CSV files (with delimiter and header inference), JSON (arrays and nested objects with automatic flattening), plain text, Markdown, log files, Notion databases (via Notion CLI), and Google Drive documents (exported as CSV or text via the gws CLI). You can point at a single file or an entire directory -- the system processes each source and merges results where appropriate.
Tips & Patterns
- Start with default mode for quick turnaround, then re-run with
--teamwhen you need a presentation-quality report. - Point at a directory when your data lives across multiple files -- the system merges results and reports which files were included.
- Use templates for recurring reports (weekly status, monthly reviews) to maintain a consistent format your stakeholders recognize.
- Business context files in
_infrastructure/context/active/are loaded automatically. If you havebusiness-info.mdandstrategy.mdset up, reports use your company terminology and align recommendations with your current strategy. - The report generator learns your preferences over time. If you consistently prefer certain chart types or section lengths, the Intelligence Engine adapts.