OpenClaw Doctor (macOS)
v1.0.0macOS Gateway 24/7 watchdog with 4-layer health checks and auto-repair. Monitors: L1 process alive, L2 HTTP port, L3 WebSocket communication (1006 detection)...
⭐ 0· 104·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
medium confidencePurpose & Capability
The name/description align with the included script: it checks process, HTTP, WebSocket errors, prevents sleep, and attempts restarts. The script calls an 'openclaw' CLI and uses launchctl/launch agents, which is reasonable for a Gateway watchdog. Minor mismatch: registry metadata in the skill header said no required binaries, but the SKILL.md metadata and the script expect commands like curl, pgrep, launchctl, caffeinate and the openclaw CLI — the openclaw binary is not declared in the registry 'Required binaries' list.
Instruction Scope
SKILL.md instructions stay within watchdog/setup duties (copy script, set env vars, install LaunchAgent). The script reads local logs (/tmp/openclaw/openclaw-YYYY-MM-DD.log) and writes logs to ~/.openclaw/logs — these are expected for debugging but worth reviewing. It also invokes 'openclaw doctor --repair' and 'openclaw gateway start' which can perform non-trivial repair actions; users should confirm what those commands do on their system before enabling automated repairs. SKILL.md references a 'references/ai.openclaw.doctor.plist' to edit, but that file is not present in the bundle.
Install Mechanism
This is an instruction-only skill with no install spec or remote downloads. The only persistent change it instructs is adding a LaunchAgent and copying the provided script to ~/.openclaw — low install-surface and no external code fetches in the package.
Credentials
No credentials or external network endpoints are requested. The script probes only a local gateway URL (default http://127.0.0.1:18789) and manipulates local services and files. Environment variables offered for customization are optional and limited to service names/URL.
Persistence & Privilege
The skill asks the user to install a LaunchAgent (normal for a watchdog) which grants persistent background execution. 'always: false' and autonomous invocation of the skill by the agent are default. This persistence is proportional for a 24/7 watchdog but means the script will run periodically and may restart services — verify you want that behavior.
Assessment
Before installing: (1) Inspect the provided scripts yourself (they are included) and confirm you trust the 'openclaw' CLI that the script invokes — the script will call openclaw doctor/gateway start which can restart services. (2) The SKILL.md mentions a references/ai.openclaw.doctor.plist but that file is not present in the package — do not copy or run an unknown plist; request the missing plist or create a LaunchAgent yourself following the documented fields. (3) Confirm service names (OPENCLAW_DOCTOR_SERVICE_NAME and OPENCLAW_DOCTOR_NODE_SERVICE) match your system's launchd labels; mismatches could cause failed restarts. (4) Note the script reads local logs (/tmp/openclaw/...) and writes to ~/.openclaw/logs; ensure those paths are acceptable and do not contain sensitive information you don't want aggregated. (5) If you enable the LaunchAgent, expect periodic automatic restarts and caffeinate (prevents sleep) — if you prefer manual control, run the script manually for testing first. If you want higher assurance, request the missing plist and explicit declaration of required binaries (including the openclaw CLI) from the skill author before installation.Like a lobster shell, security has layers — review code before you run it.
latestvk9741d6tnt1tkt6s4n554vrcb183ev5w
License
MIT-0
Free to use, modify, and redistribute. No attribution required.