report（report 工具）

Create Investigation/Analysis Report

Overview

This skill creates a structured report summarizing investigation and analysis work in GitHub Flavored Markdown (GFM) format. It follows a consistent template that organizes findings, recommendations, and conclusions in a professional format.

Workflow

1. Confirm Report Details (Claude Code only)

When running in Claude Code, use AskUserQuestion tool to confirm the following before creating the report.

Auto Mode handling: If the session is in Auto Mode (continuous, autonomous execution), skip the AskUserQuestion calls and apply these defaults: Type = Investigation (or the type the user already named in their request), Location = docs/reports/. Surface the chosen values in a one-line note in the final summary so the user can correct them. Do not block on user input in Auto Mode.

Report Type:

• Question: "What type of report are you creating?"
• Header: "Type"
• Options:

- "Investigation" - Error investigation, debugging session documentation - "Analysis" - Performance analysis, code analysis, data analysis - "Audit" - Security audit, compliance review - "Review" - Code review summary, architecture review

• The selected type will be used in the report metadata

Save Location:

• Question: "Where should the report be saved?"
• Header: "Location"
• Options:

- "docs/reports/ (Recommended)" - Standard location for reports - "docs/" - General documentation directory - "Project root" - Save in the root directory

2. Determine File Location and Name

Based on the confirmed details, determine the file name and save location following these conventions:

File name format:

{topic}-{YYYY-MM-DD}-{HHmm}.md

Save location priority:

1. docs/reports/ directory (recommended)
2. docs/ directory
3. Project root

Naming conventions:

• Topic part should concisely represent the investigation/analysis content

- Examples: error-investigation, performance-analysis, security-audit

• Date format: YYYY-MM-DD (e.g., 2025-11-13)
• Time format: HHmm in 24-hour notation (e.g., 1430)
• All lowercase, words separated by hyphens

Examples:

docs/reports/api-error-investigation-2025-11-13-1430.md
docs/reports/database-performance-analysis-2025-11-13-0915.md
docs/reports/security-vulnerability-audit-2025-11-13-1620.md

3. Verify and Create Directory

If the save destination directory does not exist, create it:

mkdir -p docs/reports

4. Create Report

Create the report following this standard structure:

Required section structure:

# {Report Title}

**Created**: YYYY-MM-DD HH:mm
**Author**: Claude Code
**Type**: {Investigation/Analysis/Audit/Review}

## Summary

{Brief explanation of the report's purpose and investigation content}

## Background

{Why this investigation/analysis was performed}

## Investigation/Analysis Details

{Detailed content}

## Findings

{Important discoveries or issues}

## Recommendations

{Future actions, proposed fixes, etc.}

## Conclusion

{Summary}

## Additional Information

{Referenced files, related documents, etc.}

Formatting requirements:

• Write in GitHub Flavored Markdown format
• Add appropriate language specification to code blocks
• Use ```diff for diff displays
• Actively use tables
• Write file paths as links: [filename.ts](path/to/filename.ts)
• Use ## or ### to create section hierarchy

5. Save File and Confirm

1. Save the report with the determined file name
2. After saving, explicitly display the file path
3. Display report summary (number of sections, character count, etc.)

Output Example

📝 Report created

Location: docs/reports/api-error-investigation-2025-11-13-1430.md
Size: 3,245 characters
Sections: 7
Created: 2025-11-13 14:30

Use Cases

• Error investigation reports
• Performance analysis documentation
• Security audit findings
• Code review summaries
• Architecture decision records
• Debugging session documentation