stripe-webhook

$npx mdskill add aAAaqwq/AGI-Super-Team/stripe-webhook

Deploy and manage Stripe webhook workers for payment notifications.

  • Automates TG alerts and CRM updates from checkout events.
  • Requires Node.js, wrangler CLI, and GitHub API access.
  • Executes logic based on Stripe webhook signing secrets.
  • Delivers results via Telegram messages and GitHub commits.

SKILL.md

.github/skills/stripe-webhookView on GitHub ↗
---
name: stripe-webhook
description: Manage AI Kitchen Pro Stripe webhook — deploy, test, switch test/prod, view logs
---
# Stripe Webhook (AI Kitchen Pro)

> Cloudflare Worker that receives Stripe checkout.session.completed events, sends TG notifications, and updates CRM via GitHub API.

## When to use

- Deploy or redeploy the webhook worker
- Switch between test and prod Stripe mode
- Test the payment notification flow
- Debug webhook failures
- View worker logs

## Dependencies

- External: Node.js, wrangler CLI (`npm i -g wrangler`), Cloudflare account
- Other skills: `git-workflow` (for commits)

## Paths

| What | Path |
|------|------|
| Worker code | `$WEBHOOK_PATH/src/index.js` |
| Wrangler config | `$WEBHOOK_PATH/wrangler.toml` |
| Secrets (.env) | `$WEBHOOK_PATH/.env` |
| GitHub repo | `your-org/your-webhook` |

## Architecture

```
Stripe checkout.session.completed
    → Cloudflare Worker (your-webhook.your-project.workers.dev/stripe)
        → 1. TG notification to Ivan (@your_pay_bot)
        → 2. CRM update (people.csv + activities.csv via GitHub API)
        → 3. Email (via Stripe receipt, no custom code)
```

## Secrets (Cloudflare Worker)

| Secret | Description |
|--------|-------------|
| `STRIPE_WEBHOOK_SECRET` | Signing secret (whsec_...) — different for test vs prod |
| `TELEGRAM_BOT_TOKEN` | @your_pay_bot token |
| `TELEGRAM_CHAT_ID` | Your Telegram user ID |
| `GITHUB_TOKEN` | PAT with repo write access |

## How to execute

### Deploy

```bash
cd $WEBHOOK_PATH
npx wrangler deploy
```

### Switch to TEST mode

```bash
# Read test secret from .env
grep STRIPE_TEST_WEBHOOK_SECRET $WEBHOOK_PATH/.env
# Set it
echo "<test_whsec>" | npx wrangler secret put STRIPE_WEBHOOK_SECRET
```

### Switch to PROD mode

```bash
grep STRIPE_LIVE_WEBHOOK_SECRET $WEBHOOK_PATH/.env
echo "<prod_whsec>" | npx wrangler secret put STRIPE_WEBHOOK_SECRET
```

### Send test webhook (simulate payment)

```python
import hmac, hashlib, time, json, urllib.request

secret = "<STRIPE_WEBHOOK_SECRET>"  # test or prod
payload = json.dumps({
    "id": "evt_test_123",
    "type": "checkout.session.completed",
    "data": {
        "object": {
            "id": "cs_test_abc",
            "object": "checkout.session",
            "amount_total": 10100,  # $101
            "currency": "usd",
            "customer_details": {"email": "[email protected]", "name": "Test User"},
            "customer_email": "[email protected]",
            "custom_fields": [{"key": "telegram_username", "text": {"value": "test_user"}}],
            "metadata": {"product_id": "prod_XXXXXXXXXXXX"}
        }
    }
})

timestamp = str(int(time.time()))
signature = hmac.new(secret.encode(), f"{timestamp}.{payload}".encode(), hashlib.sha256).hexdigest()

req = urllib.request.Request(
    "https://your-webhook.your-project.workers.dev/stripe",
    data=payload.encode(),
    headers={
        "Content-Type": "application/json",
        "Stripe-Signature": f"t={timestamp},v1={signature}",
        "User-Agent": "Stripe/1.0",
    },
    method="POST",
)
resp = urllib.request.urlopen(req, timeout=15)
print(resp.read().decode())
```

### View live logs

```bash
cd $WEBHOOK_PATH
npx wrangler tail
```

## Stripe Products

| Product ID (PROD) | Product ID (TEST) | Name | Price |
|---|---|---|---|
| prod_XXXXXXXXXXXX | prod_TEST_XXXXXXX | Tier 1 Basic | $101/mo |
| prod_YYYYYYYYYYYY | prod_TEST_YYYYYYY | Tier 2 Premium | $500 |
| prod_ZZZZZZZZZZZZ | prod_TEST_ZZZZZZZ | Tier 3 Enterprise | $1000 |

## Webhook Endpoints

| Mode | Endpoint ID | Signing Secret |
|------|-------------|----------------|
| PROD | `<YOUR_PROD_ENDPOINT_ID>` | `<YOUR_PROD_WEBHOOK_SECRET>` |
| TEST | `<YOUR_TEST_ENDPOINT_ID>` | `<YOUR_TEST_WEBHOOK_SECRET>` |

## Tier detection logic

Worker matches tier by: 1) `metadata.product_id` → TIER_MAP, 2) fallback by amount ($101/$500/$1000).

## Limitations

- CRM update goes directly to `main` branch (no PR) — acceptable for webhook automation
- No retry logic — Stripe retries failed webhooks automatically
- `custom_fields` telegram_username depends on Stripe checkout form having this field configured
- GitHub OAuth token (gho_) may expire — if CRM updates fail, refresh via `gh auth login`

## Troubleshooting

| Problem | Solution |
|---------|----------|
| TG notification not arriving | Check TELEGRAM_BOT_TOKEN, ensure /start was sent to bot |
| CRM update failing | Check GITHUB_TOKEN validity (`gh auth status`), check UTF-8 in CSV |
| Invalid signature (400) | Wrong STRIPE_WEBHOOK_SECRET — check test vs prod mode |
| SSL handshake failure | New subdomain — wait 2-5 min for cert provisioning |

## Related skills

- `payment-tracker-run` — payment follow-up automation
- `log-activity` — manual CRM activity logging

More from aAAaqwq/AGI-Super-Team

SkillDescription
a-fund-monitor监控 A 股基金实时估值与盘后净值,自动判断交易日并生成提醒或分析。
add-leadAdd company/person/relationship to CRM
ads-agentAI-агент для управления Facebook рекламой. Вызывай для анализа, оптимизации, создания кампаний и отчётов.
afrexai-compliance-auditRun internal compliance audits against major governance and security
afrexai-personal-financeComplete personal finance system — budgeting, debt payoff, investing, tax optimization, net worth tracking, and financial independence planning. Use when managing money, building wealth, paying off debt, planning retirement, or optimizing taxes. Zero dependencies.
after-salesUse when managing post-purchase experience, building customer loyalty, or increasing repeat purchases
agent-contactsAI agent contacts — add, list, remove MCP contacts. Use when someone gives an agent URL, or when you need to view/remove contacts.
agent-model-switcher批量查看和切换子 agent 的模型配置,用于统一调整多 agent 的 provider/model 设置。
agent-patternsFormat SPAWN REQUEST messages to launch parallel agents, generate structured agent status reports, and define communication protocols within the sprint system. Use when the user needs to coordinate multiple agents, format spawn requests, produce agent reports, or establish inter-agent communication patterns.
agent-task-confirmUse when confirming whether a dispatched agent task was actually received, activated, and progressing after sessions_send or other task handoff actions.