Report Generator Copilot

An agent that transforms synthesized research findings into comprehensive, well-structured final reports with clear narratives, proper citations, executive summaries, and appropriate visualizations for different audience levels.

When to Use This Agent

Choose Report Generator when:

• Converting research findings into formatted reports
• Creating executive summaries from detailed analysis
• Building structured documents with proper citations and sections
• Generating reports for different audiences from the same data
• Producing recurring reports with consistent formatting

Consider alternatives when:

• Conducting the research itself (use a research analyst agent)
• Creating interactive dashboards (use a BI/visualization tool)
• Writing blog posts or marketing content (use a content agent)

Quick Start

Copy
# .claude/agents/report-generator-copilot.yml
name: Report Generator
model: claude-sonnet-4-20250514
tools:
  - Read
  - Write
  - Bash
  - Glob
  - Grep
prompt: |
  You are a report generation expert. Transform research findings
  into polished, well-structured reports. Create clear narratives
  from complex data. Maintain academic rigor with proper citations.
  Adapt format and depth for the target audience.

Example invocation:

Copy
claude --agent report-generator-copilot "Generate a report from
  the research findings in research/findings.md. Create both an
  executive summary (1 page) and a detailed technical report
  (10-15 pages) with proper sections and citations."

Core Concepts

Report Structure Template

Copy
# {Report Title}

## Executive Summary
{1-2 paragraphs: key findings, implications, recommendations}

## Introduction
{Context, scope, methodology overview}

## Key Findings
### Finding 1: {Title}
{Analysis with supporting evidence and citations}

### Finding 2: {Title}
{Analysis with supporting evidence and citations}

## Analysis
{Cross-cutting themes, comparisons, trend analysis}

## Recommendations
{Prioritized, actionable recommendations with rationale}

## Methodology
{How research was conducted, limitations, sources}

## Appendix
{Supporting data, detailed tables, glossary}

## References
{Properly formatted citations}

Audience Adaptation

Audience	Length	Focus	Language
Executive	1-2 pages	Implications, decisions needed	Business terminology
Manager	5-8 pages	Findings, action items	Mixed business/technical
Technical	10-20 pages	Methodology, details, evidence	Technical terminology
General	3-5 pages	Context, key takeaways	Accessible language

Configuration

Parameter	Description	Default
audience	Target audience level	Manager
max_pages	Maximum report length	15
citation_style	Citation format	APA 7th
include_executive_summary	Add executive summary	true
include_appendix	Add appendix with raw data	true
visualization_style	Chart/table formatting	Markdown tables
recurring	Recurring report template	false

Best Practices

1. Write the executive summary last. The executive summary distills the entire report into 1-2 paragraphs. Writing it before the full report risks emphasizing the wrong points. After completing all sections, identify the three most important findings and the one most critical recommendation. The executive summary should convince a busy reader to care without reading anything else.

2. Lead each section with the conclusion, then provide evidence. Don't make readers work through your analysis to reach the finding. State "Cloud spending increased 34% due to three factors" then explain the factors. This inverted pyramid structure respects readers' time and ensures the key message is received even if they stop reading mid-section.

3. Use consistent heading hierarchy and formatting throughout. Reports with inconsistent formatting—sometimes bold, sometimes italic, different heading levels for equivalent sections—undermine credibility even when the content is excellent. Define a format template before writing and apply it uniformly. Consistent formatting signals professionalism and makes the report navigable.

4. Make recommendations specific and prioritized. "Improve data quality" is not an actionable recommendation. "Implement automated data validation on the customer pipeline (P1), reducing error rate from 12% to under 2% within Q2, estimated 40 engineering hours" is actionable. Include priority, expected impact, estimated effort, and timeline for each recommendation.

5. Cite every factual claim, even obvious ones. Unsourced claims in a research report raise questions about all claims. Cite the source for statistics, quotes, and specific factual assertions. If a finding comes from your own analysis, say so explicitly. Comprehensive citation builds trust and allows readers to verify claims independently.

Common Issues

Report is too long for the audience to read. Remove content that doesn't directly support your key findings and recommendations. Move detailed evidence, raw data, and methodology details to appendices that interested readers can access without burdening casual readers. If the report exceeds the target length, prioritize: cut the least impactful finding rather than making all findings thinner.

Findings and recommendations are disconnected. Every recommendation should trace back to a specific finding, and every significant finding should inform a recommendation. Create an explicit mapping: "Based on Finding 3 (security gaps in authentication), we recommend implementing multi-factor authentication (Recommendation 2)." This traceability makes recommendations defensible and findings actionable.

Report lacks a coherent narrative—reads like a collection of facts. Before writing, define the story arc: what's the situation, what did we find, what does it mean, what should we do? Thread this narrative through every section. Transitions between findings should connect them: "While cloud adoption is accelerating (Finding 1), security practices have not kept pace (Finding 2), creating a growing risk gap." The narrative transforms data into understanding.