Automations that run while you sleep.
The whole point of self-hosting an agent is that it works when you don't. This scales you from a few starter workflows to a quiet machine of scheduled jobs and event triggers — paired with the habit that makes 24/7 autonomy trustworthy: always being able to answer "what did my agent do last night?"
Scheduled jobs, done like an operator.
OpenClaw's scheduler runs any task on any rhythm — the craft is jobs that fail gracefully and stay observable. For each rule, pick the operator's version.
Overnight work goes to files; mornings get one consolidated push. Your phone stays quiet, nothing is lost — and silence always means success, never mystery.
Webhooks are an inbound door.
A webhook lets another service trigger a task the moment something happens — form submitted, repo pushed, sensor fired. But an inbound door inherits every Lesson-3 rule plus its own. Two calls:
Read your agent's diary.
If you can't answer "what did it do last night," you don't have automation — you have hope. Here's a week of your agent's activity log. Tap the entries that deserve a second look.
You can have the agent email you a weekly self-report ("jobs run, failures, anything outside the workspace, new domains, spend vs last week — flag anything unusual first"). Handy — but a compromised agent's diary can lie. The weekly skim of the real logs and the provider's billing page (which the agent can't edit) is what makes the self-report trustworthy. Defense in layers.
A mature setup is almost boring.
Six months in, the well-run box is quiet: a 7am briefing, four watchers, a nightly digest to files, one webhook from a form, a Sunday self-report — every job bounded, every failure loud, logs skimmed weekly over coffee. That boredom is the achievement. The demo-reel agents doing wild unsupervised tasks end up in scan reports; the quiet ones compound value for years.
Build your overnight layer
Add the nightly-digest job and the Sunday self-report this week. Then do your first real log skim and answer the three questions out loud — did it do what I scheduled, did it do anything I didn't, is the spend curve flat? That ritual, five minutes a week, is the entire difference between owning an agent and merely hosting one.
What you can do now
- Write cron jobs that are single-purpose, bounded, staggered, and loud about failure
- Route overnight output to files and consolidate into one morning push
- Receive webhooks through your private tunnel and treat every payload as data, not instructions
- Run the weekly log skim: scheduled vs actual, anything unscheduled, spend curve flat
- Set up an agent self-report — and know why it doesn't replace the real logs