OpenClaw & Hermes

Cronable isn’t a replacement for OpenClaw or Hermes — it’s the operations layer above them. Those agents are excellent at doing the work; Cronable is where you schedule it, chain it, secure it and watch it. The rule of thumb:

Chat to author. Canvas to operate. Talking to an agent is a great way to create a job. A visual DAG with run history is a far better way to live with fifty of them.

There are three ways to combine them.

1. Run an agent as a job

Drive an OpenClaw or Hermes agent from a Cronable job, the same way Cronable runs Claude or Codex — on a schedule or a trigger, with retries, timeouts, secret-masked logs, and its output flowing into the next step of the chain.

# jobs/research/daily-brief.yaml
id: daily-brief
group: research
type: hermes            # or: openclaw
schedule:
  - kind: daily
    time: "07:00"
prompt: |
  Summarise what changed in our space overnight and
  post the three things worth my attention to Slack.
onFailure: alert-oncall

The exact fields mirror the job’s Inspector form — see the Job reference.

2. Reuse their skills

Both ecosystems publish reusable skills. Point a Cronable job at one and run it on a schedule, with the full run history and retries you’d expect — the skill does the work, Cronable owns the when, the what-next and the what-happened.

3. Drive Cronable from the agent’s chat

Install the cronable skill in your OpenClaw or Hermes agent and it can hand recurring or critical work to Cronable by talking to your daemon. Ask the agent to “run this reliably every morning,” and it creates the job in Cronable — where it becomes a durable node on the canvas instead of a line in a chat thread that scrolls away.

Because the daemon binds to localhost, the agent reaches it the same way the dashboard does — same-origin and authenticated — and inbound events arrive through the webhook relay, so nothing is exposed.

Why route through Cronable at all?

• You can see it. Every run is on a canvas, colour-coded by state, with the masked log a click away — instead of asking the agent what it did and trusting the answer.
• It’s shared. A DAG is something your whole team (or your client) can look at; a chat is yours.
• It’s durable. Schedules, durable waits, retries, timeouts, cooldowns and failure handlers — the operational plumbing a chat loop doesn’t give you.

See Where Cronable fits for the fuller comparison.