--- title: Agentic Code Quality | daily.dev description: Argues that as AI coding agents generate enormous volumes of code, traditional human code review can't scale, so software quality must instead be enforced... canonical: https://daily.dev/posts/agentic-code-quality-7iuckhife twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Agentic Code Quality | daily.dev og:description: Argues that as AI coding agents generate enormous volumes of code, traditional human code review can't scale, so software quality must instead be enforced... og:url: https://daily.dev/posts/agentic-code-quality-7iuckhife og:image: https://api.daily.dev/og/posts/7iuCkhIfe.png og:image:alt: Agentic Code Quality 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. # Agentic Code Quality **[Addy Osmani](https://daily.dev/sources/addy)** · [@addyosmani](https://daily.dev/addyosmani) · 9 min read · 56 upvotes · 6 comments ## Summary Argues that as AI coding agents generate enormous volumes of code, traditional human code review can't scale, so software quality must instead be enforced through constraints built into the harness, environment, and pipeline around agents. These 'quality gates' include unit/property/mutation testing, complexity metrics, linting-enforced architecture rules, and compiler/CI back-pressure applied throughout the loop rather than at the end. The piece frames quality as a multidimensional set of signals (correctness, maintainability, performance, security, comprehensibility) requiring deliberate trade-offs between innovation speed and rigor, with human judgment reserved for nuanced, subjective decisions when automated guardrails fail. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How should code quality be enforced when AI agents generate most of the code? Quality shifts from manual code review to constraints embedded in the harness, environment, and CI pipeline. These 'quality gates' include unit, property, and acceptance tests, mutation testing to catch sneaky bugs, and metrics like cyclomatic complexity. Human reviewers are reserved for cases where automated guardrails fail, since human attention is scarce and machine-speed changes overwhelm manual review. _daily.dev surfaces practical approaches for teams building constraint-driven quality gates around coding agents._ ### What should a team do if verification can't keep up with the volume of AI agent generated changes? Three options exist: scale up the verification system to handle more capacity, reduce the rate at which agents generate changes so checks can catch up, or lower the quality bar so verification pushes back less. Teams should be ready to apply all three depending on context, and may even loosen constraints in lower-risk areas to boost throughput while tightening them where quality matters most. _Teams tuning agent throughput against quality bars can track these trade-off strategies on daily.dev._ ## Community discussion Top comments from developers on daily.dev. **@michaelbarsinai** · 4 upvotes > All good points. We deal with this using (deterministic) model-based testing, which gives us known and controllable coverage of the system being validated. **@rizzdev** · 1 upvotes > Three scaling options, and option 3 is lowering the bar. Guillermo skip-reading the low-stakes PRs is just calendar math. Mutation testing scoring the tests the same agent just wrote still prints a number, like a report card the suite filled in for itself **@kartiknvj** · 0 upvotes > Pushing quality gates into the loop instead of the end matches what finally worked for me. The piece leans on unit and mutation tests, but for agent output I had to add behavioral evals as a gate too, since code compiles and passes types while still giving wrong answers. Where do you draw the line on which checks block the merge versus just warn? **@agustinbarrientos** · 0 upvotes > Mutation reports should break results down by agent and change type. Each workflow also needs a seeded mutation it has failed before. A second survivor should open a regression issue automatically. ## Similar posts on daily.dev - [Can agentic coding raise the quality bar?](https://daily.dev/posts/can-agentic-coding-raise-the-quality-bar--a8pvbbdom) · Luca Palmieri · 2 upvotes · 0 comments - [Agentic Code Review](https://daily.dev/posts/agentic-code-review-0ue8fcsa1) · Addy Osmani · 146 upvotes · 14 comments - [QCon chat: Is agentic AI killing continuous integration?](https://daily.dev/posts/qcon-chat-is-agentic-ai-killing-continuous-integration--a6ctqyqkx) · The New Stack · 0 upvotes · 0 comments - [Can QA Reignite its Purpose in the Agentic Code Generation Era?](https://daily.dev/posts/can-qa-reignite-its-purpose-in-the-agentic-code-generation-era--5vvbqrmfl) · DevOps.com · 0 upvotes · 0 comments - [Human Maintainable Code in the Age of AI Agents](https://daily.dev/posts/human-maintainable-code-in-the-age-of-ai-agents-ilzajja2s) · Atomic Spin · 12 upvotes · 1 comments --- Tags: [#testing](https://daily.dev/tags/testing), [#architecture](https://daily.dev/tags/architecture), [#cicd](https://daily.dev/tags/cicd), [#ai-coding](https://daily.dev/tags/ai-coding) [View this post on daily.dev](https://daily.dev/posts/agentic-code-quality-7iuckhife) ```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":"Agentic Code Quality","url":"https://daily.dev/posts/agentic-code-quality-7iuckhife","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/agentic-code-quality-7iuckhife"},"datePublished":"2026-08-23T20:02:12.893Z","dateModified":"2026-09-14T07:00:21.978Z","description":"Argues that as AI coding agents generate enormous volumes of code, traditional human code review can't scale, so software quality must instead be enforced...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/3e55197f8bf421f2a24a9b6347fc827e?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/3e55197f8bf421f2a24a9b6347fc827e?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Addy Osmani","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":"Person","name":"Addy Osmani","url":"https://daily.dev/addyosmani","image":"https://avatars2.githubusercontent.com/u/110953?v=4","interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"EndorseAction"},"userInteractionCount":5510}},"commentCount":6,"discussionUrl":"https://daily.dev/posts/agentic-code-quality-7iuckhife","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":56},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":6}],"keywords":"testing,architecture,cicd,ai-coding","timeRequired":"PT9M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Addy Osmani","item":"https://daily.dev/sources/addy"},{"@type":"ListItem","position":3,"name":"Agentic Code Quality"}]} {"@context":"https://schema.org","@type":"WebPage","@id":"https://daily.dev/posts/agentic-code-quality-7iuckhife","comment":[{"@type":"Comment","text":"All good points. We deal with this using (deterministic) model-based testing, which gives us known and controllable coverage of the system being validated.","datePublished":"2026-08-24T17:51:28.982Z","url":"https://daily.dev/posts/7iuCkhIfe#c-hT3cjsjgW","author":{"@type":"Person","name":"Michael Bar-Sinai","url":"https://daily.dev/michaelbarsinai","image":"https://avatars.githubusercontent.com/u/1938468?v=4"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":4}},{"@type":"Comment","text":"Three scaling options, and option 3 is lowering the bar. Guillermo skip-reading the low-stakes PRs is just calendar math. Mutation testing scoring the tests the same agent just wrote still prints a number, like a report card the suite filled in for itself","datePublished":"2026-08-25T15:55:19.356Z","url":"https://daily.dev/posts/7iuCkhIfe#c-lhgE4NRPa","author":{"@type":"Person","name":"Andrew","url":"https://daily.dev/rizzdev","image":"https://media.daily.dev/image/upload/s--35vRfXAA--/f_auto/v1785413727/avatars/avatar_fQ7ttCKtPuu6tDvcgQXv5?_a=BAMAMicg0"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1}},{"@type":"Comment","text":"Pushing quality gates into the loop instead of the end matches what finally worked for me. The piece leans on unit and mutation tests, but for agent output I had to add behavioral evals as a gate too, since code compiles and passes types while still giving wrong answers. Where do you draw the line on which checks block the merge versus just warn?","datePublished":"2026-08-26T18:45:10.857Z","url":"https://daily.dev/posts/7iuCkhIfe#c-a4Olfe5X2","author":{"@type":"Person","name":"kartik-nvjk","url":"https://daily.dev/kartiknvj","image":"https://media.daily.dev/image/upload/s--3gGgsVCw--/f_auto/v1781456774/avatars/avatar_TvTVeiMdkRCqWUDullFmy?_a=BAMAMiWQ0"}},{"@type":"Comment","text":"Mutation reports should break results down by agent and change type. Each workflow also needs a seeded mutation it has failed before. A second survivor should open a regression issue automatically.","datePublished":"2026-08-26T02:37:32.005Z","url":"https://daily.dev/posts/7iuCkhIfe#c-haHn9ENNu","author":{"@type":"Person","name":"Agustin Barrientos","url":"https://daily.dev/agustinbarrientos","image":"https://media.daily.dev/image/upload/s--5ayxQnqn--/f_auto/v1788281802/avatars/avatar_wQYYVe5Tbj0NJ7C7qPoa8?_a=BAMAMicg0"}}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/agentic-code-quality-7iuckhife#faq","mainEntity":[{"@type":"Question","name":"How should code quality be enforced when AI agents generate most of the code?","acceptedAnswer":{"@type":"Answer","text":"Quality shifts from manual code review to constraints embedded in the harness, environment, and CI pipeline. These 'quality gates' include unit, property, and acceptance tests, mutation testing to catch sneaky bugs, and metrics like cyclomatic complexity. Human reviewers are reserved for cases where automated guardrails fail, since human attention is scarce and machine-speed changes overwhelm manual review. daily.dev surfaces practical approaches for teams building constraint-driven quality gates around coding agents."}},{"@type":"Question","name":"What should a team do if verification can't keep up with the volume of AI agent generated changes?","acceptedAnswer":{"@type":"Answer","text":"Three options exist: scale up the verification system to handle more capacity, reduce the rate at which agents generate changes so checks can catch up, or lower the quality bar so verification pushes back less. Teams should be ready to apply all three depending on context, and may even loosen constraints in lower-risk areas to boost throughput while tightening them where quality matters most. Teams tuning agent throughput against quality bars can track these trade-off strategies on daily.dev."}}]} ```