Report Generate

Generate a polished Markdown report from data sources. Operate in one of two modes depending on arguments.

Parse Arguments

Extract these flags from $ARGUMENTS:

• --team (boolean, default: false) — activate full 5-agent pipeline mode
• --data=PATH (string, optional) — path to data source file or directory. If not provided, ask the user what data to analyze.
• --output=PATH (string, default: ./report-output/) — output directory for generated report

Business Context (Optional)

Check if context files exist at _infrastructure/context/active/. If the directory contains .md files, read business-info.md, strategy.md, and current-data.md. Use this context to personalize output (e.g., prioritize known clients, use correct terminology, align with current strategy). If files don't exist, skip silently.

Step 0: Memory Context

Read the context-injection skill at _infrastructure/memory/context-injection/SKILL.md. Query for memories relevant to the current input (company, contacts, topics detected in arguments). If memories are returned, incorporate them into your working context for this execution.

Mode 1: Default (Single-Agent Quick Report)

When --team is NOT present:

1. Read ALL 5 skills:
   • ${CLAUDE_PLUGIN_ROOT}/skills/report/data-extraction/SKILL.md
   • ${CLAUDE_PLUGIN_ROOT}/skills/report/data-analysis/SKILL.md
   • ${CLAUDE_PLUGIN_ROOT}/skills/report/report-writing/SKILL.md
   • ${CLAUDE_PLUGIN_ROOT}/skills/report/chart-generation/SKILL.md
   • ${CLAUDE_PLUGIN_ROOT}/skills/report/executive-summary/SKILL.md
2. If --data is provided, read the data source. Auto-detect format (CSV, JSON, text).
3. If no --data, ask the user for their data source or report requirements.
4. Perform analysis using the data-analysis skill rules.
5. Generate a concise report (3-5 pages) with:
   • Executive summary
   • Key findings with supporting metrics
   • 1-2 Mermaid charts for key data visualizations
   • Recommendations
6. Write the report to --output as a Markdown file.
7. Present a summary to the user with the output path.

Output Format (Default Mode)

## Report Generated

**Output**: ./report-output/report-2026-02-23.md
**Data Sources**: 1 file (sales-q4.csv)
**Sections**: Executive Summary, Key Findings, Analysis, Recommendations
**Charts**: 2 Mermaid diagrams embedded
**Word Count**: ~2,800 words

### Executive Summary Preview
[First 2-3 sentences of the executive summary]

Mode 2: Team Pipeline (--team)

When --team IS present:

1. Read the pipeline configuration at ${CLAUDE_PLUGIN_ROOT}/agents/report/config.json.
2. Execute the full 5-agent pipeline sequentially:
   • Research Agent → gather and extract data from all sources
   • Analysis Agent → process, analyze, identify trends
   • Writing Agent → generate report prose with executive summary
   • Formatting Agent → add Mermaid charts, format tables, write output file
   • QA Agent → review for accuracy and consistency (recommend-only)
3. Each agent reads its definition from ${CLAUDE_PLUGIN_ROOT}/agents/report/.
4. Pass structured JSON output from each agent as input to the next.
5. Present the final pipeline report with:
   • Report output path
   • Page count and section summary
   • Charts generated
   • QA notes and flags (if any)
   • Notion DB entry in "Founder OS HQ - Reports" with Type="Business Report" (if Notion configured)

Data Source Detection

Detect the data source type and handle accordingly:

• .csv files → Parse as comma-separated values. Extract headers as column names, detect numeric vs. text columns, handle quoted fields and escaped commas. Summarize row count and column types.
• .json files → Parse as JSON. Support both array-of-objects and nested structures. Flatten nested keys using dot notation for analysis. Detect numeric fields for statistical analysis.
• .txt / .md / .log files → Read as plain text. Extract structured data via pattern matching (tables, key-value pairs, bullet lists). For log files, parse timestamps and group by time intervals.
• Notion database → Query via Notion MCP (if configured). Use notion-search to find the database by name, then fetch all pages. Extract property values as columns for analysis.
• Directory path → Scan for all supported file types (.csv, .json, .txt, .md, .log). Process each file individually, then merge results. Report which files were included and any that were skipped.

Final: Memory Update

Read the pattern-detection skill at _infrastructure/memory/pattern-detection/SKILL.md. Log this execution as an observation with: plugin name, primary action performed, key entities (companies, contacts), and output summary. Check for emerging patterns per the detection rules. If a memory reaches the adaptation threshold, append the notification to the output.

Error Handling

• If no data source provided and user doesn't specify one: halt and ask.
• If data file doesn't exist: report error with path.
• If MCP servers unavailable in team mode: warn but continue with filesystem-only.
• If Notion MCP not configured: skip Notion DB logging, output to filesystem only.
• If output directory doesn't exist: create it.

Usage Examples

/founder-os:report:generate --data=sales-q4.csv
/founder-os:report:generate --data=./data/ --output=./reports/
/founder-os:report:generate --team --data=metrics.json
/founder-os:report:generate --team --data=sales-q4.csv --output=./q4-report/
/founder-os:report:generate                              # Interactive: asks for data source