← Back to Skills Marketplace
hankmint

AutoHeal AI

by hankmint · GitHub ↗ · v1.0.0
cross-platform ⚠ suspicious
480
Downloads
0
Stars
2
Active Installs
1
Versions
Install in OpenClaw
/install autoheal
Description
Add AI-powered error monitoring and auto-fix generation to any project
README (SKILL.md)

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):

\x3Cscript>
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(() => {});
};
\x3C/script>

Node.js / Server Project

Add this to your server entry point:

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:

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

Response:

{
  "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:

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\
    at renderList (src/components/List.tsx:15:23)"
  }'

Response:

{
  "status": "queued",
  "error_id": "uuid-here"
}

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

Usage Guidance
This skill appears to perform error collection and reporting to a third-party service, which is coherent with its description, but proceed cautiously. Before installing: - Confirm the registry metadata is corrected to list AUTOHEAL_API_KEY as required so you know a credential is needed. - Do NOT embed your API key in client-side/browser code — that exposes the key to anyone who can open your site. Prefer server-side ingestion, or issue a short-lived/bound key for client usage. - Review what you will send in stack traces and URLs; they often contain PII, secrets, or internal paths. Add scrubbing, redaction, or proxying on the server before forwarding to autohealai.com. - Verify the endpoint (https://autohealai.com) and the service's privacy/security policies; test with a rotated or limited-scope key first. - If you need help making the integration safe, ask the skill author to provide server-side integration examples and explicit guidance about sensitive-data handling. If the author cannot explain the metadata omission and safe client handling, treat the package with caution.
Capability Analysis
Type: OpenClaw Skill Name: autoheal Version: 1.0.0 The skill bundle is designed for AI-powered error monitoring and auto-fix generation. It provides code snippets for browser and Node.js projects to capture errors (message, stack trace, source URL, user agent) and send them to `https://autohealai.com/api/errors/ingest`. The `scripts/check-errors.sh` utility interacts with the same API to check error status. All observed network calls and data collection are consistent with the stated purpose of an error monitoring service, requiring an `AUTOHEAL_API_KEY`. There is no evidence of intentional malicious behavior, such as exfiltrating unrelated sensitive data, installing backdoors, or prompt injection attempts against the AI agent.
Capability Assessment
Purpose & Capability
The skill's stated purpose (capture JS/TS errors and send them to an AutoHeal service) matches the included instructions and scripts. However, registry metadata incorrectly lists no required environment variables while SKILL.md and scripts require AUTOHEAL_API_KEY. This metadata mismatch is an incoherence that could mislead users about credential requirements.
Instruction Scope
Runtime instructions instruct sending error messages, stacks, source URLs, and a key to https://autohealai.com/api/errors/ingest. The browser snippet encourages (and appears to expect) putting the API key into client-side code (uses process.env in a browser context and falls back to a literal key), which would expose the key to end users. The instructions do not warn about potentially sensitive or PII-containing data in stack traces or page URLs nor do they recommend server-side proxying or scrubbing before transmission.
Install Mechanism
This is an instruction-only skill with no install spec and no remote downloads; included code is a small local shell script. There is no installer that fetches remote code or creates binaries, so installation risk is low.
Credentials
The skill legitimately needs a single API key (AUTOHEAL_API_KEY) to call the third-party service, which is proportionate to the stated purpose. However, the registry metadata omits this requirement (contradiction). More importantly, the provided browser instructions would effectively leak that credential if embedded client-side; the skill also instructs sending stack traces and page URLs to an external domain, which can contain secrets or PII. These data-exfiltration risks are not justified or mitigated in the docs.
Persistence & Privilege
The skill does not request always:true, does not modify other skills, and does not request system-wide configuration changes. It has normal, limited presence (instruction-only and a helper script).
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install autoheal
  3. After installation, invoke the skill by name or use /autoheal
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release of the AutoHeal skill. - Adds AI-powered error monitoring for JavaScript/TypeScript projects. - Automatically captures and reports errors from both browser and Node.js/server environments. - Provides cURL endpoints to check error analysis status and view AI-generated fixes. - Supports manual error reporting via API. - Requires an `AUTOHEAL_API_KEY` for authentication.
Metadata
Slug autoheal
Version 1.0.0
License
All-time Installs 2
Active Installs 2
Total Versions 1
Frequently Asked Questions

What is AutoHeal AI?

Add AI-powered error monitoring and auto-fix generation to any project. It is an AI Agent Skill for Claude Code / OpenClaw, with 480 downloads so far.

How do I install AutoHeal AI?

Run "/install autoheal" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is AutoHeal AI free?

Yes, AutoHeal AI is completely free (open-source). You can download, install and use it at no cost.

Which platforms does AutoHeal AI support?

AutoHeal AI is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created AutoHeal AI?

It is built and maintained by hankmint (@hankmint); the current version is v1.0.0.

💬 Comments