--- title: How I Develop .NET Apps With AI and Still Trust the Code description: A developer describes a set of rules, skills and hooks for reviewing AI-generated .NET code built with Uno Platform. The uno-write-spec skill forces unresolved... canonical: https://daily.dev/posts/how-i-develop-net-apps-with-ai-and-still-trust-the-code-l5rryqbfy twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: How I Develop .NET Apps With AI and Still Trust the Code | daily.dev og:description: A developer describes a set of rules, skills and hooks for reviewing AI-generated .NET code built with Uno Platform. The uno-write-spec skill forces unresolved... og:url: https://daily.dev/posts/how-i-develop-net-apps-with-ai-and-still-trust-the-code-l5rryqbfy og:image: https://api.daily.dev/og/posts/l5rrYQBfy.png og:image:alt: How I Develop .NET Apps With AI and Still Trust the Code og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # How I Develop .NET Apps With AI and Still Trust the Code **[Uno Platform](https://daily.dev/sources/uno-platform)** · 9 min read · 0 upvotes · 1 comments ## Summary A developer describes a set of rules, skills and hooks for reviewing AI-generated .NET code built with Uno Platform. The uno-write-spec skill forces unresolved requirements into writing before any code is produced. uno-break-it runs a separate QA agent through failure scenarios like double-fired navigation and stale responses, each logged with a severity and screenshot. A change-cost probe in uno-audit rates proposed changes as Contained, Scattered, or Unguarded; one example found validation copy restated in the page and both .resw files with no test coverage. A stop hook blocks 'done' claims in an Uno Platform repository unless a completeness report is newer than the last commit. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How can I verify an AI coding agent's work is actually done before trusting it? Require evidence rather than a claim: a build, tests and lint passing with zero warnings on the touched target, a unit test for any new business rule, and runtime confirmation for UI changes. For 'done' or 'ready to demo' claims specifically, a full completeness audit should be saved to a dedicated report file, and a stop hook can block the agent's claim if that report is missing or older than the last source commit. _Developers building AI review workflows can find these evidence patterns and similar engineering practices on daily.dev._ ### What kinds of edge cases should a QA agent test for in a form with a save button? A separate QA agent should try scenarios like double-fired navigation, navigating mid-load, a spammed toggle, forced errors, and an older network response arriving after a newer one. Each finding it surfaces gets a severity rating and a screenshot, giving reviewers concrete evidence of failures beyond what a quick manual click-through would reveal. _Teams hardening AI-assisted features against edge cases can track these QA patterns on daily.dev._ ### How do I assess how risky a code change will be before making it? Run a change-cost probe that takes the app's main feature plus a few plausible change requests, then for each one lists the files it would touch, the test that would catch a mistake, and a verdict of Contained, Scattered, or Unguarded. In one real example, a validation rule change was rated 'Scattered' because the rule text was duplicated across a page and two resource files with no test covering that copy. _Engineers weighing refactor risk before shipping can follow these change-cost evaluation techniques on daily.dev._ ## Community discussion Top comments from developers on daily.dev. **@bickov** · 0 upvotes > Does uno-break-it save each screenshot as a file and put the path in the report? If the QA agent runs as a Claude Code subagent, the main session never sees images it looked at, only the text it hands back. Were the severity notes alone enough to fix the stale response cases? ## Similar posts on daily.dev - [96% of developers don’t trust AI code: Here’s a step toward the fix](https://daily.dev/posts/96-of-developers-don-t-trust-ai-code-here-s-a-step-toward-the-fix-m3vixpqtq) · The New Stack · 2 upvotes · 0 comments - [I trust AI code that crashes more than AI code that works, and 66% of developers seem to agree](https://daily.dev/posts/i-trust-ai-code-that-crashes-more-than-ai-code-that-works-and-66-of-developers-seem-to-agree-yswlwmidg) · XDA Developers · 0 upvotes · 0 comments - [How to Evaluate AI Code Quality: A Practical Guide for Engineers](https://daily.dev/posts/how-to-evaluate-ai-code-quality-a-practical-guide-for-engineers-rbtcmfynl) · freeCodeCamp · 0 upvotes · 0 comments - [Beyond Code Generation: How AI Is Reshaping Modern Software Delivery](https://daily.dev/posts/beyond-code-generation-how-ai-is-reshaping-modern-software-delivery-olilqg6hs) · SitePoint · 2 upvotes · 0 comments - [Stop Pretending You Trust the Code Your AI Just Wrote](https://daily.dev/posts/stop-pretending-you-trust-the-code-your-ai-just-wrote-z5gqyktus) · Medium · 0 upvotes · 0 comments --- Tags: [#testing](https://daily.dev/tags/testing), [#.net](https://daily.dev/tags/.net), [#ai-coding](https://daily.dev/tags/ai-coding), [#code-review](https://daily.dev/tags/code-review) [View this post on daily.dev](https://daily.dev/posts/how-i-develop-net-apps-with-ai-and-still-trust-the-code-l5rryqbfy) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"How I Develop .NET Apps With AI and Still Trust the Code","url":"https://daily.dev/posts/how-i-develop-net-apps-with-ai-and-still-trust-the-code-l5rryqbfy","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/how-i-develop-net-apps-with-ai-and-still-trust-the-code-l5rryqbfy"},"datePublished":"2026-10-26T19:05:12.555Z","dateModified":"2026-10-26T19:23:20.174Z","description":"A developer describes a set of rules, skills and hooks for reviewing AI-generated .NET code built with Uno Platform. The uno-write-spec skill forces unresolved...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/0da871bb5108e4709a90153a8fbf1704?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/0da871bb5108e4709a90153a8fbf1704?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Uno Platform","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Uno Platform","logo":"https://media.daily.dev/image/upload/s--04Uc7--c--/f_auto,q_auto/v1783840777/logos/uno-platform?_a=BAMAMicg0","url":"https://daily.dev/sources/uno-platform"},"commentCount":1,"discussionUrl":"https://daily.dev/posts/how-i-develop-net-apps-with-ai-and-still-trust-the-code-l5rryqbfy","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":1}],"keywords":"testing,.net,ai-coding,code-review","timeRequired":"PT9M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Uno Platform","item":"https://daily.dev/sources/uno-platform"},{"@type":"ListItem","position":3,"name":"How I Develop .NET Apps With AI and Still Trust the Code"}]} {"@context":"https://schema.org","@type":"WebPage","@id":"https://daily.dev/posts/how-i-develop-net-apps-with-ai-and-still-trust-the-code-l5rryqbfy","comment":[{"@type":"Comment","text":"Does uno-break-it save each screenshot as a file and put the path in the report? If the QA agent runs as a Claude Code subagent, the main session never sees images it looked at, only the text it hands back. Were the severity notes alone enough to fix the stale response cases?","datePublished":"2026-10-27T08:15:27.842Z","url":"https://daily.dev/posts/l5rrYQBfy#c-s5LJcrHwq","author":{"@type":"Person","name":"ab","url":"https://daily.dev/bickov","image":"https://lh3.googleusercontent.com/a/ACg8ocKwOmJ31UWqM4ZXB5ms-yRdFR4BYzyb7FPBRpg1O6DV0MIAF2EP=s96-c"}}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/how-i-develop-net-apps-with-ai-and-still-trust-the-code-l5rryqbfy#faq","mainEntity":[{"@type":"Question","name":"How can I verify an AI coding agent's work is actually done before trusting it?","acceptedAnswer":{"@type":"Answer","text":"Require evidence rather than a claim: a build, tests and lint passing with zero warnings on the touched target, a unit test for any new business rule, and runtime confirmation for UI changes. For 'done' or 'ready to demo' claims specifically, a full completeness audit should be saved to a dedicated report file, and a stop hook can block the agent's claim if that report is missing or older than the last source commit. Developers building AI review workflows can find these evidence patterns and similar engineering practices on daily.dev."}},{"@type":"Question","name":"What kinds of edge cases should a QA agent test for in a form with a save button?","acceptedAnswer":{"@type":"Answer","text":"A separate QA agent should try scenarios like double-fired navigation, navigating mid-load, a spammed toggle, forced errors, and an older network response arriving after a newer one. Each finding it surfaces gets a severity rating and a screenshot, giving reviewers concrete evidence of failures beyond what a quick manual click-through would reveal. Teams hardening AI-assisted features against edge cases can track these QA patterns on daily.dev."}},{"@type":"Question","name":"How do I assess how risky a code change will be before making it?","acceptedAnswer":{"@type":"Answer","text":"Run a change-cost probe that takes the app's main feature plus a few plausible change requests, then for each one lists the files it would touch, the test that would catch a mistake, and a verdict of Contained, Scattered, or Unguarded. In one real example, a validation rule change was rated 'Scattered' because the rule text was duplicated across a page and two resource files with no test covering that copy. Engineers weighing refactor risk before shipping can follow these change-cost evaluation techniques on daily.dev."}}]} ```