README.md

---
title: AI Vidya Daily Report Agent
emoji: 📋
colorFrom: indigo
colorTo: blue
sdk: gradio
sdk_version: 5.9.1
app_file: app.py
pinned: false
---

# AI Vidya — Daily Report Evaluation Agent

Reads **#ai-vidya** on Slack, checks who submitted their EOD report,
evaluates quality with Claude AI, and posts results to **#intern-updates**.

## Setup (5 minutes)

### 1. Create the HuggingFace Space
- Go to huggingface.co → New Space
- SDK: **Gradio**, Visibility: **Private**
- Upload `app.py`, `requirements.txt`, `README.md`

### 2. Add Secrets
In your Space → **Settings → Variables and Secrets**:

| Secret name | Value |
|---|---|
| `SLACK_BOT_TOKEN` | `xoxb-your-bot-token` |
| `ANTHROPIC_API_KEY` | `sk-ant-your-key` |

### 3. Create the Slack Bot
1. Go to [api.slack.com/apps](https://api.slack.com/apps) → Create New App
2. **OAuth & Permissions** → Bot Token Scopes, add:
   - `channels:history` — read #ai-vidya messages
   - `channels:read` — find channel IDs
   - `chat:write` — post to #intern-updates
   - `users:read` — look up intern names
3. Install to workspace → copy **Bot User OAuth Token**
4. Invite bot to both channels:
   - `/invite @your-bot-name` in **#ai-vidya**
   - `/invite @your-bot-name` in **#intern-updates**

### 4. Get Slack User IDs for Interns
In Slack: click any intern's profile → ⋮ (more) → **Copy member ID**
It looks like: `U06T83LP86P`

Paste them into the intern registry JSON in the UI.

### 5. Schedule Daily at 11 PM IST

**GitHub Actions (free, recommended):**

Create `.github/workflows/daily_check.yml`:
```yaml
name: Daily Report Check
on:
  schedule:
    - cron: '30 17 * * 1-5'  # 11:00 PM IST (Mon–Fri) = 17:30 UTC
jobs:
  trigger:
    runs-on: ubuntu-latest
    steps:
      - name: Trigger HF Space
        run: |
          curl -X POST "${{ secrets.HF_SPACE_URL }}/run/predict" \
            -H "Content-Type: application/json" \
            -d '{"data": [""]}'
```

Add `HF_SPACE_URL` as a GitHub secret — format: `https://your-username-your-space-name.hf.space`

## What it does

1. Reads all messages in **#ai-vidya** posted today
2. Checks each registered intern — submitted or missed
3. Evaluates submitted reports with Claude:
   - 🟢 **Good** — specific technical content, real blocker, shows thinking
   - 🟡 **Weak** — vague, missing blocker, just a status update
   - 🔴 **Invalid** — one liner, link only, "done", "working on it"
4. Posts a full summary to **#intern-updates**
5. @mentions each intern who missed with a warning message