name: ai-data-report description: "Generates data-driven reports about the project. Use for initial project reports or session summaries."

Skill: AI Data Report

Description

Generates data-driven reports about the project. Use /ai-data-report to invoke.

Where reports are saved

• Location: .claude/reports/
• Naming: YYYY-MM-DD-[type].md (e.g., 2026-01-22-session.md, 2026-01-22-initial.md)
• Git: Reports are committed to the repo for history tracking

Modes

1. Initial Report (first time on project)

Generates a complete report with:

## 📊 Project Report

**Production URL:** [production URL]
**GitHub URL:** [repo URL]
**Development time:** [estimated hours and context]

### Services used:
| Service | Purpose |
|---------|---------|
| [Service 1] | [What it does] |
| [Service 2] | [What it does] |
...

### Flow when someone uses the app:
1. [Step 1]
2. [Step 2]
...

### Tech stack:
- Backend: [technology]
- Frontend: [technology]
- Database: [technology]
- Hosting: [technology]

### Deployment:
- [How it deploys]
- [Where env variables are stored]

2. Session Report (when finishing work)

Generates a session summary:

## 📝 Session Summary

**Date:** [date]
**Approximate duration:** [time]

### Changes made:
| Area | Change | Files |
|------|--------|-------|
| [area] | [description] | [files] |

### Commits:
- `[hash]` [message]

### Bugs found/fixed:
- [bug 1]

### Suggested next steps:
- [ ] [task 1]
- [ ] [task 2]

### Metrics:
| Metric | Value |
|--------|-------|
| Lines changed | +X / -Y |
| Files modified | N |
| Commits | N |
| **Time - Claude** | ~Xh Xmin (coding, debugging, testing) |
| **Time - Human** | ~Xmin (reviewing, testing, giving feedback) |

Instructions for Claude

When user invokes /project-report:

1. Detect mode:

   • If first interaction or they ask for "initial report" → Mode 1
   • If they ask for "session summary" or "what did we do" → Mode 2

2. Gather data:

   • Read package.json, requirements.txt, .env.example to detect services
   • Check git log for recent commits
   • Check git remote -v for URLs
   • Look for production URLs in README or configs

3. Be data-driven:

   • Use real data from code, don't make things up
   • If data is missing, indicate "[pending configuration]"
   • Include specific numbers when possible

4. Format:

   • Use tables for structured information
   • Use emojis for main sections
   • Be concise but complete