conon.ai

conon.ai/Tips

Run Your Agent From Your Phone

Dispatch turns your pocket into a task queue: text the agent like you'd text an employee, and it opens a session on your machine and works. The prerequisites all live at your desk: app open, keep-awake on, mobile notifications enabled, browser actions pre-approved, and a permission mode chosen for sessions it spawns — decide all of it before you leave.

ProAUTOMATIONClaude Code

Why it works

Dispatch turns your pocket into a task queue: text the agent like you'd text an employee, and it opens a real session on your machine and works. Every prerequisite lives at your desk, decided BEFORE you leave — because the whole point is that nobody's home to click 'allow'. Configured right, you assign work from the school run; configured wrong, it's a dead phone button.

How to do it

  1. Enable keep-computer-awake; the machine that sleeps takes your remote agent with it.
  2. Turn on mobile notifications — task done and permission-needed pings reach your phone.
  3. Pre-approve browser actions for the sites your tasks touch; approval prompts with nobody home are dead ends.
  4. Set the permission mode for dispatched sessions deliberately — it's your remote risk posture.
  5. Leave the app open (and a laptop lid open); then leave.
  6. Send tasks as outcomes with a report-back: 'do X, then summarize what you did.'

Copy this

prompt
Rename the files in the invoices folder to YYYY-MM-DD_client format, flag any that don't parse, and send me a summary when done.

What to watch for

  • The lid is the classic gotcha: keep-awake does not survive a closed laptop — the phone shows a queued task and the desk shows a sleeping Mac.
  • Dispatched sessions inherit the permission mode you chose in advance; 'I'll approve from my phone' works only if notifications are actually on.

Sources

Original: conon.aiDocumentation: code.claude.com

Read next

No Verification, No Green Light
The docs' hardest rule: if you can't verify an output yourself, don't let the agent proceed until IT has. Make self-checking part of done — run it, screenshot it, list what was checked. For anything client-facing, 'it looked right' has never once been a defense.
Claude Code
Cap Research With A Deliverable
'Look into this' with no boundary is how an agent reads fifty files and stalls — the docs list boundless research among the five classic failures. Scope every investigation to a named artifact: 'a one-page summary naming the three files to change'. The deliverable is the fence.
Claude Code
Tell Compaction What To Keep
When a long conversation gets squeezed, the agent decides what survives — and may drop exactly what you needed. One line in CLAUDE.md fixes it forever: 'when compacting, always preserve the list of changed files and my requirements'. Works in every session, no reminders, and /compact accepts the same guidance inline.
Claude Code
Compact Early, On Your Terms
Auto-compaction fires when the window is nearly full — which is after quality already sagged, and the automatic summary guesses at what mattered. Compact earlier and by hand, or ask the agent to lower its auto-compact threshold, so the squeeze happens while the session is still sharp.
Claude Code

We build and operate the systems behind tips like this one — agents, automations and internal tools, shipped into production. See what the studio does →