Mcp Health Monitor
v1.0.0Auto-monitor MCP servers and AI services with health checks, auto-restart on failure, and Telegram alerts
⭐ 0· 49·0 current·0 all-time
by@reikys
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description, README, SKILL.md, and the included script all describe the same behavior: HTTP and process checks, optional Telegram alerts, and macOS launchctl restarts. There are no unexplained environment variables, external services, or binaries requested beyond those needed (curl/pgrep/launchctl which are expected for this task).
Instruction Scope
Runtime instructions and the script stay within monitoring/restart/alert scope. Notable points: the script sources an ENV_FILE (default $HOME/.env) which will execute any shell constructs present there (the script uses `source`), so the file must be restricted to simple VAR=VALUE entries and owned by the user. Process checks use `pgrep -f` with user-supplied patterns — overly-broad patterns can match unintended processes (false positives/negatives). Alerts are sent only to Telegram's API (https://api.telegram.org) as expected; no other external endpoints are contacted.
Install Mechanism
No install spec — instruction-only with an included shell script. This is low-risk from an installation perspective: nothing is downloaded from remote URLs and nothing is automatically written to system locations by an installer. The user is instructed to copy the script and create a LaunchAgent or cron entry manually.
Credentials
No required secrets are declared in registry metadata; the script optionally uses TELEGRAM_BOT_TOKEN and TELEGRAM_CHAT_ID for alerts, which is proportionate. Caveat: the script sources the ENV_FILE which could execute commands if the file contains shell statements rather than simple assignments; treat the .env file as executable content and protect it accordingly.
Persistence & Privilege
The skill does not request 'always: true' or other elevated platform privileges. The installation instructions recommend creating a user LaunchAgent or cron entry to run periodically — this gives persistent scheduling (intended for monitoring) but is a normal and expected level of persistence for a health-check tool.
Assessment
This skill appears to do what it says: local health checks, optional Telegram alerts, and macOS launchctl restarts. Before installing: (1) open and review scripts/healthcheck.sh and the SERVICES array and adjust patterns so pgrep -f won't accidentally match unrelated processes; (2) if you use a .env file, ensure it contains only VAR=VALUE lines and is owned and writable only by your user (sourcing the file executes shell code); (3) test the script manually before enabling a scheduled LaunchAgent/cron job; (4) verify the Telegram bot token and chat ID are correct and that alerts go only to your intended chat; (5) on Linux, adapt the restart commands to systemctl instead of launchctl. If you are comfortable with these checks, the skill is coherent with its stated purpose.Like a lobster shell, security has layers — review code before you run it.
latestvk9724nbwgr371tva6v2jxnjfgh83x35e
License
MIT-0
Free to use, modify, and redistribute. No attribution required.