Skeptical Report Reader - Claude Code Template

A Claude Code workspace template for critical analysis of reports and lengthy documents. This template provides a library of analysis commands and a structured workflow for extracting insights from reports.

Derived from: Skeptical-Report-Reader - a NotebookLM prompt library for critical report analysis.

Quick Start

1. Clone or use this template for your analysis project
2. Run /init to configure the workspace with your project context
3. Place your report(s) in /input
4. Run /preprocessing:convert-to-markdown to prepare documents
5. Run analysis commands like /summarization:comprehensive
6. Find outputs in /output

First-Time Setup

After cloning, run the /init command. This will ask you about:

• What reports you're analyzing (single report or a cluster)
• Why you're interested in these reports
• What specific questions you want answered
• Areas requiring extra scrutiny

Your answers are saved to context.md and inform all subsequent analyses.

Preprocessing

Before running analysis commands, convert documents to markdown for cleaner, more reliable parsing:

# Recommended: Install docling for PDF/document conversion
pip install docling

# Alternative: markdownify for HTML content
pip install markdownify

Then run /preprocessing:convert-to-markdown to convert all documents in /input. Converted files are saved to /input/markdown/.

Available Commands

Setup & Customization

Command	Description
/init	Initialize workspace with project context
/add-command	Create custom slash commands for your needs

Preprocessing

Command	Description
/preprocessing:convert-to-markdown	Convert PDFs, DOCX, HTML to markdown

Triage - AI as Judge

Use these to quickly assess whether a report merits deeper analysis:

Command	Description
/triage:quick-scan	Rapid assessment—what's interesting about this report?
/triage:original-thinking	Does this report contain genuinely new ideas?
/triage:worth-the-read	Verdict: is this worth the reader's time?
/triage:bias-check	Evaluate for bias and conflicts of interest

Analysis - Arguments & Evidence

Deep dives into how the report makes its case:

Command	Description
/analysis:main-arguments	Core thesis, supporting arguments, evidence strength
/analysis:case-studies	Analysis of case studies used as evidence
/analysis:methodology-review	Critical evaluation of research methods

Statistics - Quantitative Extraction

Focus on the numbers:

Command	Description
/statistics:key-statistics	All major statistics organized thematically
/statistics:noteworthy-findings	Unexpected or novel findings with page refs
/statistics:data-quality	Assess reliability of data presented

Summarization

Executive summaries and quotable content:

Command	Description
/summarization:exec-summary	Critical executive summary with strengths/weaknesses
/summarization:comprehensive	Full 7-section analysis
/summarization:key-quotes	Quotable passages and soundbites

Workflow

1. /init                              → Configure workspace with your context
2. Add reports to /input              → PDFs, DOCX, HTML, or other documents
3. /preprocessing:convert-to-markdown → Convert to markdown for cleaner parsing
4. /triage:worth-the-read             → Quick assessment: should you dig deeper?
5. Run analysis commands              → Choose based on your needs
6. Review /output                     → Analyses saved with descriptive filenames
7. /add-command                       → Create project-specific commands as needed

Philosophy

This template embodies a skeptical approach to report analysis:

• Critical evaluation over passive summarization: Don't just regurgitate—assess credibility, methodology, and originality
• Respect the reader's time: Quickly determine if a report is worth reading
• Add analytical value: Consider what's missing, identify biases, evaluate evidence strength
• Be honest: If a report is corporate filler, say so

Directory Structure

.
├── CLAUDE.md              # Claude Code instructions, persona, and repo map
├── README.md              # This file
├── context.md             # [Created by /init] Your project-specific context
├── input/                 # Place reports here
│   └── markdown/          # [Created by preprocessing] Converted documents
├── output/                # Generated analyses saved here
└── .claude/
    └── commands/
        ├── init.md
        ├── add-command.md
        ├── preprocessing/
        │   └── convert-to-markdown.md
        ├── triage/
        │   ├── quick-scan.md
        │   ├── original-thinking.md
        │   ├── worth-the-read.md
        │   └── bias-check.md
        ├── analysis/
        │   ├── main-arguments.md
        │   ├── case-studies.md
        │   └── methodology-review.md
        ├── statistics/
        │   ├── key-statistics.md
        │   ├── noteworthy-findings.md
        │   └── data-quality.md
        └── summarization/
            ├── exec-summary.md
            ├── comprehensive.md
            └── key-quotes.md

Creating Custom Commands

Use /add-command to create analysis commands tailored to your specific needs. Examples:

• A command focused on extracting competitor mentions
• A command that identifies funding sources and potential conflicts
• A command that compares findings across multiple reports

Custom commands are saved to the appropriate subfolder and become part of your project's command library.

Use Cases

• Research: Quickly assess whether academic papers or industry reports merit deep reading
• Competitive Intelligence: Extract key data points from competitor reports
• Due Diligence: Critically evaluate claims in investment or partnership materials
• Literature Review: Process multiple reports and extract common themes
• Policy Analysis: Identify evidence strength and gaps in policy documents

---

For more Claude Code projects, visit my Claude Code Repos Index.