<?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: Dong Hong</title>
    <description>The latest articles on DEV Community by Dong Hong (@donghp).</description>
    <link>https://dev.to/donghp</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%2F4253312%2Fcd9a8af7-6df3-4f77-a49a-f06bc8c4a00c.png</url>
      <title>DEV Community: Dong Hong</title>
      <link>https://dev.to/donghp</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/donghp"/>
    <language>en</language>
    <item>
      <title>Vibe Coding Is Easy. Building Reliable Software Is Not.</title>
      <dc:creator>Dong Hong</dc:creator>
      <pubDate>Thu, 19 Nov 2026 08:26:38 +0000</pubDate>
      <link>https://dev.to/donghp/vibe-coding-is-easy-building-reliable-software-is-not-2932</link>
      <guid>https://dev.to/donghp/vibe-coding-is-easy-building-reliable-software-is-not-2932</guid>
      <description>&lt;p&gt;AI coding tools have changed the speed of software creation forever. With a few prompts, anyone can scaffold a full-stack app, wire up an API, and ship a working prototype in an afternoon.&lt;/p&gt;

&lt;p&gt;That is &lt;strong&gt;vibe coding&lt;/strong&gt;—and for rapid prototyping, it is magical.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;The problem starts on Day 3.&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;When a project grows beyond a few files, unbounded prompting hits a wall:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;The AI agent forgets architectural constraints established 20 messages ago.&lt;/li&gt;
&lt;li&gt;Fixing one bug silently breaks an unrelated module.&lt;/li&gt;
&lt;li&gt;You spend more time debugging AI-generated regressions than writing the feature yourself.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;&lt;strong&gt;Chat history is not project memory.&lt;/strong&gt; If you want to build reliable, production-grade software with AI, you have to move from &lt;em&gt;prompt-centric&lt;/em&gt; chaos to &lt;strong&gt;Project-Centric Engineering&lt;/strong&gt;.&lt;/p&gt;




&lt;h2&gt;
  
  
  Why AI Coding Agents Lose Context at Scale
&lt;/h2&gt;

&lt;p&gt;Large Language Models do not "understand" your system architecture the way a senior engineer does. They predict tokens based on the active context window. When your context is scattered across long chat threads, the signal-to-noise ratio collapses.&lt;/p&gt;

&lt;p&gt;To keep human engineers in control while letting AI execute at full speed, every project needs explicit engineering boundaries before code is written.&lt;/p&gt;




&lt;h2&gt;
  
  
  Two Artifacts That Immediately Stabilize AI Coding
&lt;/h2&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%2Folyr2n0g9dvwsmpy9g2f.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%2Folyr2n0g9dvwsmpy9g2f.png" alt="Context Manifest Architecture" width="800" height="1200"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Instead of writing longer, more desperate prompts, introduce two lightweight, machine-readable files into your repository root.&lt;/p&gt;

&lt;h3&gt;
  
  
  1. The Context Manifest (&lt;code&gt;context-manifest.json&lt;/code&gt;)
&lt;/h3&gt;

&lt;p&gt;A Context Manifest replaces fragile chat memory with a single authoritative source of truth that the AI reads at the start of every session:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight json"&gt;&lt;code&gt;&lt;span class="p"&gt;{&lt;/span&gt;
  &lt;span class="nl"&gt;"project"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"payment-gateway-service"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
  &lt;span class="nl"&gt;"architecture"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;
    &lt;span class="nl"&gt;"pattern"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"Hexagonal / Ports and Adapters"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
    &lt;span class="nl"&gt;"invariants"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="p"&gt;[&lt;/span&gt;
      &lt;span class="s2"&gt;"Domain layer must have zero external framework imports"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
      &lt;span class="s2"&gt;"All database mutations must pass through repository interfaces"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
      &lt;span class="s2"&gt;"Never expose internal entity IDs directly in REST responses"&lt;/span&gt;
    &lt;span class="p"&gt;]&lt;/span&gt;
  &lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
  &lt;span class="nl"&gt;"active_modules"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="p"&gt;[&lt;/span&gt;
    &lt;span class="s2"&gt;"src/domain/transactions/"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
    &lt;span class="s2"&gt;"src/adapters/stripe/"&lt;/span&gt;
  &lt;span class="p"&gt;]&lt;/span&gt;
&lt;span class="p"&gt;}&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h3&gt;
  
  
  2. The Task Contract (&lt;code&gt;task-contract.json&lt;/code&gt;)
&lt;/h3&gt;

&lt;p&gt;Your AI agent cannot read your mind. Before asking it to write or refactor code, lock down four boundaries:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;
&lt;strong&gt;Objective:&lt;/strong&gt; What exact outcome must change?&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Scope:&lt;/strong&gt; Which specific files are allowed to be modified?&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Constraints:&lt;/strong&gt; What existing behavior or interfaces must remain untouched?&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Acceptance:&lt;/strong&gt; What test command or verification evidence proves the task is complete?
&lt;/li&gt;
&lt;/ol&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight json"&gt;&lt;code&gt;&lt;span class="p"&gt;{&lt;/span&gt;
  &lt;span class="nl"&gt;"task_id"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"FEAT-042"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
  &lt;span class="nl"&gt;"objective"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"Add idempotency key validation to webhook handler"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
  &lt;span class="nl"&gt;"scope"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;
    &lt;span class="nl"&gt;"allowed_files"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="p"&gt;[&lt;/span&gt;
      &lt;span class="s2"&gt;"src/adapters/stripe/webhookHandler.ts"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
      &lt;span class="s2"&gt;"tests/adapters/stripe/webhookHandler.test.ts"&lt;/span&gt;
    &lt;span class="p"&gt;]&lt;/span&gt;
  &lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
  &lt;span class="nl"&gt;"constraints"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="p"&gt;[&lt;/span&gt;
    &lt;span class="s2"&gt;"Do not modify database schema or existing retry backoff logic"&lt;/span&gt;
  &lt;span class="p"&gt;]&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
  &lt;span class="nl"&gt;"acceptance_criteria"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="p"&gt;[&lt;/span&gt;
    &lt;span class="s2"&gt;"npm test -- webhookHandler.test.ts passes with 100% branch coverage"&lt;/span&gt;
  &lt;span class="p"&gt;]&lt;/span&gt;
&lt;span class="p"&gt;}&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;






&lt;h2&gt;
  
  
  Creative Freedom Requires Engineering Boundaries
&lt;/h2&gt;

&lt;p&gt;AI can generate code at 10x speed, but &lt;strong&gt;humans must still own the architecture, trade-offs, and verification&lt;/strong&gt;. When you anchor AI agents with explicit manifests and contracts, you keep the creative flow of vibe coding without sacrificing production reliability.&lt;/p&gt;

&lt;h3&gt;
  
  
  Explore the Open Templates &amp;amp; Full Book
&lt;/h3&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Open JSON Schemas &amp;amp; Engineering Templates:&lt;/strong&gt; Inspect and fork the full reference templates on GitHub at &lt;a href="https://github.com/donghp/beyond-vibe-coding" rel="noopener noreferrer"&gt;github.com/donghp/beyond-vibe-coding&lt;/a&gt;.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Read the First 20% Free:&lt;/strong&gt; These patterns are adapted from my book, &lt;strong&gt;&lt;em&gt;Beyond Vibe Coding: A Practical Guide to Building Real-World Software with AI&lt;/em&gt;&lt;/strong&gt;. You can read the first 20% free on &lt;a href="https://play.google.com/store/books/details?id=geQaEgAAQBAJ" rel="noopener noreferrer"&gt;Google Play Books&lt;/a&gt;.&lt;/li&gt;
&lt;/ul&gt;




&lt;p&gt;&lt;strong&gt;How are you currently keeping your AI coding agents from drifting off-architecture on larger codebases?&lt;/strong&gt; Drop your workflow in the comments—I would love to compare notes!&lt;/p&gt;

</description>
      <category>ai</category>
      <category>webdev</category>
      <category>productivity</category>
      <category>programming</category>
    </item>
  </channel>
</rss>
