windows-healing-gateway

Automates monitoring and self-repair of OpenClaw Gateway on Windows using Task Scheduler, with AI diagnosis and Telegram alert notifications.

Do not install or run this skill until you can review the actual PowerShell scripts it will install and execute. Specific actions to take before trusting it: - Ask the author for the full contents of all .ps1 scripts (deploy-windows-healing.ps1, openclaw-monitor.ps1, openclaw-fix.ps1, openclaw-gateway-starter.ps1). Do not run ExecutionPolicy Bypass on unreviewed scripts. - Verify why MOONSHOT_API_KEY and CODING_PLAN_KEY are required, what services they authenticate to, and whether those services are trustworthy. Confirm these keys are necessary for the skill's stated diagnosis feature. - Confirm how Telegram tokens are used and that alerts are sent only to the configured chat ID. Prefer token storage methods you control (not embedding tokens in uploaded repos). - Confirm the repository/origin URL (skill.json points to github.com/franco-user, some docs use 'yourusername' placeholders) and prefer installing only from a trusted, reviewable repo/tagged release. - Request that the skill metadata declare required env vars so automated systems can flag sensitive requirements. - If you must test, run initial deployment in an isolated VM or disposable environment and inspect created Task Scheduler tasks and script behavior before giving it production privileges or real API keys. These inconsistencies (missing script files, undeclared sensitive env vars, hidden scripted tasks) make the package risky until audited; treat it as suspicious rather than safe.

Type: OpenClaw Skill Name: windows-healing-gateway Version: 1.0.0 The skill implements a persistence mechanism using Windows Task Scheduler (OpenClaw-Monitor-Service.xml) and executes PowerShell scripts with 'ExecutionPolicy Bypass' and hidden windows. It manages sensitive credentials, including API keys and Telegram tokens, within environment files. While the stated intent is 'self-healing' and monitoring, the core logic scripts (e.g., openclaw-monitor.ps1, openclaw-fix.ps1) are missing from the provided bundle, making the high-privilege background operations and potential for data exfiltration unverifiable.