Heartbeat Monitor

SkillDocs & knowledge

External uptime and liveness check for the bot, memory pipelines, and local API. Use when the user wants to monitor whether the agent is alive, set up a health check, get notified when the bot goes down, or debug why scheduled jobs stopped running.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the Heartbeat Monitor skill

What this skill tells your AI

The instructions your AI receives, as published by thesmokedev/taskchad-os in optional-skills/devops/heartbeat-monitor/SKILL.md and read by ahel’s review.

Check that the running stack is healthy and report a single green/red verdict.

What to probe

The framework runs a local API on port 4322 (see .claude/sections/03_* / orchestration). Probe the pieces that matter:

ComponentProbe
Local APIGET http://127.0.0.1:4322/health returns 200
Chat processprocess for run_chat.sh is alive
Memory heartbeatlast heartbeat timestamp is < 2× the interval old
Diskvault partition has free space

Run

bash optional-skills/devops/heartbeat-monitor/scripts/healthcheck.sh
# exit 0 = all green, exit 1 = something is down

The script prints a per-check table and sets a non-zero exit code on any failure, so it composes with cron / CI / an external uptime service.

Escalation

  • For local use, surface the red verdict in chat.
  • For unattended monitoring, point an external uptime service (Healthchecks.io, UptimeRobot, a cron on another host) at this script or at the /health route — a monitor running inside the same process it watches can't report its own death.
  • This skill only reads state; it does not restart anything. Pair with log-triage to explain why a check went red.

Signals

GitHub stars
23
Forks
5
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
heartbeat-monitor
Source
github.com/thesmokedev/taskchad-os