<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Rico.Deng</title>
    <description>The latest articles on DEV Community by Rico.Deng (@zakiso).</description>
    <link>https://dev.to/zakiso</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4168575%2Fd414ccd6-532b-46d0-b00e-8be8961fb0b1.jpg</url>
      <title>DEV Community: Rico.Deng</title>
      <link>https://dev.to/zakiso</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/zakiso"/>
    <language>en</language>
    <item>
      <title>Keeping a local coding-agent workflow usable from a phone</title>
      <dc:creator>Rico.Deng</dc:creator>
      <pubDate>Wed, 07 Oct 2026 10:24:05 +0000</pubDate>
      <link>https://dev.to/zakiso/keeping-a-local-coding-agent-workflow-usable-from-a-phone-16i2</link>
      <guid>https://dev.to/zakiso/keeping-a-local-coding-agent-workflow-usable-from-a-phone-16i2</guid>
      <description>&lt;p&gt;I'm the developer of BotBus, a free companion for coding agents running on a computer. I built it for the moment when an agent is still working but I'm away from the keyboard: I want to read the task context, reply to a question, or inspect the result from my phone.&lt;/p&gt;

&lt;h2&gt;
  
  
  A workflow to try
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;Install the BotBus host on the computer where your agent runs.&lt;/li&gt;
&lt;li&gt;Pair your phone with the computer using BotBus's pairing flow.&lt;/li&gt;
&lt;li&gt;Open the relevant task, read its context and continue the conversation. Where the connector supports approvals, you can respond from the phone.&lt;/li&gt;
&lt;li&gt;Inspect a result shared by the agent, such as an image, video, file or PDF.&lt;/li&gt;
&lt;li&gt;If the result needs investigation, open the computer's file browser or terminal.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;BotBus works with Codex, Claude Code, Hermes, OpenCode and DeepSeek Harness. The available actions depend on the connector and task state; seeing an agent in the list does not mean every action is available for that agent.&lt;/p&gt;

&lt;h2&gt;
  
  
  Task results and the computer workspace
&lt;/h2&gt;

&lt;p&gt;There are two useful entry points. A task's shared results let you inspect what the agent produced. The computer workspace lets you browse files and use a terminal when you need to investigate further.&lt;/p&gt;

&lt;h3&gt;
  
  
  Browse files
&lt;/h3&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fiiq41uw2icbyle99a48a.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fiiq41uw2icbyle99a48a.png" alt="BotBus computer file browser" width="800" height="1739"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h3&gt;
  
  
  Connect a terminal
&lt;/h3&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Faggytwccu57cymkw1zfv.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Faggytwccu57cymkw1zfv.png" alt="BotBus terminal on a phone" width="800" height="1739"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h3&gt;
  
  
  Open a development preview
&lt;/h3&gt;

&lt;p&gt;On macOS, you can also open development webpages from the computer and control its screen. Linux and Windows hosts currently provide files and terminals, without screen control or development previews.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fkch1vr06seizbfc7qob9.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fkch1vr06seizbfc7qob9.png" alt="BotBus development webpage preview" width="800" height="1739"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;These are current promotional screenshots with simulator/demo content. The video below uses simulator screens and composed scenes, rather than documenting a complete real-device test.&lt;/p&gt;

&lt;p&gt;  &lt;iframe src="https://www.youtube.com/embed/dv9nmE0IQ0g" width="710" height="399"&gt;
  &lt;/iframe&gt;
&lt;/p&gt;

&lt;h2&gt;
  
  
  Practical boundaries
&lt;/h2&gt;

&lt;p&gt;The agent continues to run on your computer, so the computer must stay online. Computer hosts are available for macOS, Linux and Windows. Mobile options include iPhone, Apple Watch and an Android APK from the website; the APK is not a claim of Google Play availability.&lt;/p&gt;

&lt;p&gt;Task content is end-to-end encrypted. Development previews have a separate security boundary, described in the &lt;a href="https://botbus.io/en/docs/security" rel="noopener noreferrer"&gt;security documentation&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;BotBus is free. The full app is not open source, but its &lt;a href="https://github.com/botbus-io/botbus-protocol" rel="noopener noreferrer"&gt;protocol and connector packages are public&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://botbus.io/en/?utm_source=devto&amp;amp;utm_medium=community&amp;amp;utm_campaign=botbus_launch_202610#download" rel="noopener noreferrer"&gt;Download and installation&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;I'd appreciate specific feedback on the first pairing experience and the transition from reading an agent task to inspecting its files. Which step is unclear, and what context would you need before taking action from a phone?&lt;/p&gt;

</description>
      <category>productivity</category>
      <category>showdev</category>
      <category>ai</category>
    </item>
  </channel>
</rss>
