ZappushZappush
SkillsUse CasesBenchmarkCommunitySign In
      Back to Skills
      Zjianru

      Safety Report

      Restart Guard

      @Zjianru

      Deterministic OpenClaw gateway restart with down/up state-machine verification, origin-session proactive ACK, and backward-compatible config.

      1,674Downloads
      6Installs
      4Stars
      7Versions
      API Integration4,971Legal & Compliance738

      Security Analysis

      medium confidence
      Suspicious0.08 risk

      The skill's code and instructions mostly match its restart-and-notify purpose, but it is designed to auto-run high-impact gateway restarts from natural-language triggers and it reads local env/dotfile state and external-notification secrets in ways worth reviewing before install.

      Mar 2, 202621 files4 concerns
      Purpose & Capabilitynote

      Name/description (deterministic gateway restart + notifications) aligns with the included scripts and config. Required binaries (python3, curl) and GATEWAY_AUTH_TOKEN are consistent with the HTTP/CLI notification paths and restart operations. The skill also includes multi-channel notification support (telegram/discord/slack/webhook/openclaw passthrough), which is expected for its stated delivery guarantees.

      Instruction Scopeconcern

      SKILL.md and the implementation explicitly require the agent to auto-run the full restart flow on simple natural-language triggers (e.g. "restart now") and to infer the origin session automatically. That gives the agent discretion to perform a destructive, high-impact operation without additional interactive confirmation. The code reads local files (openclaw.json, ~/.openclaw/.env) and writes local diagnostics/context files; while these are relevant to restart/notification, automatic inference and the directive "do not expose internal scripts/steps unless user explicitly asks" reduce transparency and increase risk of unintended restarts or hidden actions.

      Install Mechanismok

      This is an instruction-only skill with bundled Python scripts (no external install/download step). There is no remote URL or archive extract in the install spec — code is present in the skill bundle and runs locally. No high-risk install mechanism detected.

      Credentialsconcern

      The skill declares a single required env var (GATEWAY_AUTH_TOKEN) which is reasonable for using the gateway HTTP tool. However, notify logic will read other notification secrets (TELEGRAM_BOT_TOKEN, DISCORD_WEBHOOK_URL, SLACK_WEBHOOK_URL, RESTART_GUARD_WEBHOOK_URL, etc.) from environment or from a user dotfile (~/.openclaw/.env). The code will open and parse that dotfile to resolve keys, which can expose unrelated secrets stored there. The number/variety of optional env keys is proportional to multi-channel notifications but the implicit dotenv access and passthrough to external webhooks warrant caution.

      Persistence & Privilegenote

      always is false and the skill does not request permanent platform-wide inclusion. It spawns a detached guardian process to persist across a gateway restart (intended behavior) and the guardian exits after completion. Autonomous invocation is enabled (platform default); combined with the explicit 'must auto-run' trigger policy in SKILL.md this increases the operational blast radius, but the skill itself does not set always:true or modify other skills' configs.

      Guidance

      This skill will automatically perform gateway restarts and then try to proactively notify the originating session and external channels. Before installing or enabling it: - Be aware it is designed to auto-run on simple natural-language triggers (e.g. "restart now") and may restart a gateway without extra confirmation. Consider requiring explicit human confirmation in your deployment if you don't want one-shot voice/text triggers to restart production. - Review and limit the GATEWAY_AUTH_TOKEN and any notification tokens (Telegram/Discord/Slack/webhook) it will use. The code will also read ~/.openclaw/.env to resolve tokens — inspect that file for other secrets before allowing the skill access. - Audit the included scripts (restart.py, guardian.py, notify.py, write_context.py) in your environment and test in a staging system first. Pay attention to the notification webhook URLs and content templates to avoid accidental secret leakage. - If you want stricter safety, modify the workflow to require an explicit confirmation step before executing the auto_restart flow, or limit agent permissions so automatic invocation cannot be triggered by casual chat.

      Latest Release

      v2.2.0

      Security fixes and work directory rename

      More by @Zjianru

      Web Search Pro

      6 stars

      Ops Framework

      2 stars

      self-improving-agent

      @pskoett · 1,456 stars

      Gog

      @steipete · 672 stars

      Tavily Web Search

      @arun-8687 · 620 stars

      Find Skills

      @JimLiuxinghai · 529 stars

      Published by @Zjianru on ClawHub

      Zappush© 2026 Zappush
      HomeGuaranteeSupport

      Something feels unusual? We want to help: [email protected]