autoheal

Original source / Raw SKILL.md

---
name: autoheal
version: 1.0.0
description: Add AI-powered error monitoring and auto-fix generation to any project
requires:
  env:
    - AUTOHEAL_API_KEY
---

# AutoHeal AI Skill

AutoHeal captures production JavaScript/TypeScript errors, analyzes them with AI, and generates platform-specific fix prompts you can paste directly into your AI coding tool.

## 1. Setup AutoHeal in a Project

### Browser Project (React, Next.js, Vue, Svelte, etc.)

Add this snippet to your app's entry point (e.g., `main.tsx`, `layout.tsx`, `App.vue`):

```html
<script>
window.onerror = function(msg, source, line, col, err) {
  fetch("https://autohealai.com/api/errors/ingest", {
    method: "POST",
    headers: { "Content-Type": "application/json" },
    body: JSON.stringify({
      key: process.env.AUTOHEAL_API_KEY || "YOUR_API_KEY",
      message: msg,
      stack: err?.stack || "",
      source_url: source,
      browser: navigator.userAgent
    })
  }).catch(() => {});
};

window.onunhandledrejection = function(e) {
  const err = e.reason;
  fetch("https://autohealai.com/api/errors/ingest", {
    method: "POST",
    headers: { "Content-Type": "application/json" },
    body: JSON.stringify({
      key: process.env.AUTOHEAL_API_KEY || "YOUR_API_KEY",
      message: err?.message || String(err),
      stack: err?.stack || "",
      source_url: window.location.href,
      browser: navigator.userAgent
    })
  }).catch(() => {});
};
</script>
```

### Node.js / Server Project

Add this to your server entry point:

```javascript
process.on('uncaughtException', (err) => {
  fetch("https://autohealai.com/api/errors/ingest", {
    method: "POST",
    headers: { "Content-Type": "application/json" },
    body: JSON.stringify({
      key: process.env.AUTOHEAL_API_KEY,
      message: err.message,
      stack: err.stack || ""
    })
  }).catch(() => {});
});

process.on('unhandledRejection', (reason) => {
  const err = reason instanceof Error ? reason : new Error(String(reason));
  fetch("https://autohealai.com/api/errors/ingest", {
    method: "POST",
    headers: { "Content-Type": "application/json" },
    body: JSON.stringify({
      key: process.env.AUTOHEAL_API_KEY,
      message: err.message,
      stack: err.stack || ""
    })
  }).catch(() => {});
});
```

## 2. Check Error Status

After an error is ingested, check its analysis status:

```bash
curl -s "https://autohealai.com/api/errors/{ERROR_ID}/status" \
  -H "Authorization: Bearer $AUTOHEAL_API_KEY"
```

Response:
```json
{
  "status": "analyzed",
  "message": "Cannot read properties of undefined",
  "has_fix": true,
  "fix_id": "uuid-here",
  "dashboard_url": "https://autohealai.com/dashboard/errors/{ERROR_ID}"
}
```

Possible statuses: `new`, `analyzing`, `analyzed`, `fix_applied`, `ignored`

View the full fix with AI-generated fix prompt at the `dashboard_url`.

## 3. Report an Error Manually

Send any error directly to AutoHeal:

```bash
curl -X POST "https://autohealai.com/api/errors/ingest" \
  -H "Content-Type: application/json" \
  -d '{
    "key": "'$AUTOHEAL_API_KEY'",
    "message": "TypeError: Cannot read properties of undefined (reading '\''map'\'')",
    "stack": "TypeError: Cannot read properties of undefined\n    at renderList (src/components/List.tsx:15:23)"
  }'
```

Response:
```json
{
  "status": "queued",
  "error_id": "uuid-here"
}
```

The error will be analyzed by AI within seconds. Check status using the error_id from the response.


---

## Skill Companion Files

> Additional files collected from the skill directory layout.

### README.md

```markdown
# AutoHeal AI - OpenClaw Skill

An [OpenClaw](https://github.com/anthropics/openclaw) skill that lets AI agents add AutoHeal error monitoring to any JavaScript/TypeScript project.

## What is AutoHeal AI?

AutoHeal AI captures production JavaScript errors, analyzes them with AI, and generates platform-specific fix prompts for vibe coders using tools like Lovable, Replit, Bolt, v0, Cursor, and more.

## Getting Your API Key

1. Sign up at [autohealai.com](https://autohealai.com)
2. Go to **Settings** > **API Key**
3. Copy your `ah_...` key
4. Set it as an environment variable: `export AUTOHEAL_API_KEY=ah_your_key_here`

## Using the Skill

Once installed, you can ask your AI agent:

- "Set up AutoHeal error monitoring in this project"
- "Check the status of my AutoHeal error"
- "Report this error to AutoHeal"

The agent will use the skill instructions to integrate AutoHeal into your project.

## Helper Scripts

### check-errors.sh

Poll an error's analysis status:

```bash
# One-time check
./scripts/check-errors.sh <error_id>

# Poll until analyzed (checks every 2s, timeout 120s)
./scripts/check-errors.sh <error_id> --poll
```

## Publishing to OpenClaw

1. Fork or clone the [OpenClaw registry](https://github.com/anthropics/openclaw)
2. Copy this `SKILL.md` to the appropriate directory in the registry
3. Submit a pull request

Or reference it directly:
```
openclaw install github:hankmint/autoheal-ai/packages/openclaw-skill
```

```

### _meta.json

```json
{
  "owner": "hankmint",
  "slug": "autoheal",
  "displayName": "AutoHeal AI",
  "latest": {
    "version": "1.0.0",
    "publishedAt": 1772319543167,
    "commit": "https://github.com/openclaw/skills/commit/bba92fa80c27504c771784bfc18cad57f76ec2f4"
  },
  "history": []
}

```

### scripts/check-errors.sh

```bash
#!/usr/bin/env bash
# Check AutoHeal error status with polling
# Usage: ./check-errors.sh <error_id> [--poll]

set -euo pipefail

if [ -z "${AUTOHEAL_API_KEY:-}" ]; then
  echo "Error: AUTOHEAL_API_KEY environment variable is required"
  exit 1
fi

if [ -z "${1:-}" ]; then
  echo "Usage: $0 <error_id> [--poll]"
  exit 1
fi

ERROR_ID="$1"
POLL="${2:-}"
API_URL="https://autohealai.com/api/errors/${ERROR_ID}/status"

check_status() {
  curl -sf "$API_URL" -H "Authorization: Bearer $AUTOHEAL_API_KEY"
}

if [ "$POLL" = "--poll" ]; then
  echo "Polling error $ERROR_ID..."
  for i in $(seq 1 60); do
    RESPONSE=$(check_status) || { echo "Request failed"; exit 1; }
    STATUS=$(echo "$RESPONSE" | grep -o '"status":"[^"]*"' | cut -d'"' -f4)
    echo "  [$i] Status: $STATUS"
    if [ "$STATUS" = "analyzed" ] || [ "$STATUS" = "fix_applied" ] || [ "$STATUS" = "ignored" ]; then
      echo ""
      echo "$RESPONSE" | python3 -m json.tool 2>/dev/null || echo "$RESPONSE"
      exit 0
    fi
    sleep 2
  done
  echo "Timeout: error not yet analyzed after 120s"
  exit 1
else
  RESPONSE=$(check_status) || { echo "Request failed"; exit 1; }
  echo "$RESPONSE" | python3 -m json.tool 2>/dev/null || echo "$RESPONSE"
fi

```