--- title: The AI-Native SDLC Starts With Your Infrastructure description: Anthropic's AI-Native SDLC playbook lays out six stages for coding-agent-driven development but never specifies what an agent's tests should actually run... canonical: https://daily.dev/posts/the-ai-native-sdlc-starts-with-your-infrastructure-2fviwpotm twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: The AI-Native SDLC Starts With Your Infrastructure | daily.dev og:description: Anthropic's AI-Native SDLC playbook lays out six stages for coding-agent-driven development but never specifies what an agent's tests should actually run... og:url: https://daily.dev/posts/the-ai-native-sdlc-starts-with-your-infrastructure-2fviwpotm og:image: https://api.daily.dev/og/posts/2FViWpOtM.png og:image:alt: The AI-Native SDLC Starts With Your Infrastructure 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. # The AI-Native SDLC Starts With Your Infrastructure **[MetalBear](https://daily.dev/sources/metalbear)** · 6 min read · 1 upvotes · 1 comments ## Summary Anthropic's AI-Native SDLC playbook lays out six stages for coding-agent-driven development but never specifies what an agent's tests should actually run against. The gap matters most at Stage 4, where the agent verifies its own work: if tests run against local fakes of dependent services, a passing test only proves the code works against stale fakes, not the real cluster. mirrord is presented as the fix, letting agent code run against real staging services, queues, and databases instead of mocks, and its commercial operator lets multiple concurrent agent sessions share one staging cluster without interfering with each other via header-based traffic filtering, queue splitting, and database branching. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### Why can't an AI coding agent be trusted to judge whether its tests passing against fake services actually means the code works in production? An agent lacks the tacit knowledge a developer has about how stale a set of fakes has become, such as a mocked billing service still using an auth header that changed months ago, or a fake search endpoint that never paginates results the way the real one does. The agent only sees tests pass and reports the work finished, without knowing whether those fakes still resemble the real dependency. _daily.dev surfaces practical takes like this for developers rethinking how agent-generated code gets verified._ ### How does mirrord let AI coding agents test code against real staging services instead of local mocks? mirrord proxies a locally running process's environment variables, secrets, and network traffic to and from a shared staging cluster, so the code behaves as if deployed there without actually being deployed. This lets an agent's outbound calls hit real upstream services and see real queue contents, rather than working against fakes on its own machine or relying on documentation. _developers evaluating agent verification tooling can track approaches like mirrord's on daily.dev._ ### How does mirrord handle multiple AI coding agent sessions running against the same shared staging cluster without interfering with each other? The mirrord operator filters traffic by header so each agent session only receives its own requests, splits queues so each session gets a private slice of a shared topic, and branches databases so a session that writes does not disturb others. This lets one staging cluster serve many concurrent agent sessions instead of forcing each agent onto its own slow, expensive environment copy. _teams scaling multi-agent coding workflows can follow infrastructure patterns like this via daily.dev._ ## Community discussion Top comments from developers on daily.dev. **@leonidbugaev** · 0 upvotes > I've had the fake billing service still return 200 after the real auth header changed. The agent saw green and called it done. ## Similar posts on daily.dev - [Agentic development hinges on verification. For cloud-native software, that is a runtime problem.](https://daily.dev/posts/agentic-development-hinges-on-verification-for-cloud-native-software-that-is-a-runtime-problem--aknq5rc05) · The New Stack · 1 upvotes · 0 comments - [Testing Is the New Bottleneck for AI-Driven Software Development](https://daily.dev/posts/testing-is-the-new-bottleneck-for-ai-driven-software-development-azhhistin) · MetalBear · 0 upvotes · 0 comments - [What Happens After AI Writes the Code](https://daily.dev/posts/what-happens-after-ai-writes-the-code-nwd4gpoui) · MetalBear · 19 upvotes · 1 comments --- Tags: [#testing](https://daily.dev/tags/testing), [#kubernetes](https://daily.dev/tags/kubernetes), [#ai-agents](https://daily.dev/tags/ai-agents), [#claude-code](https://daily.dev/tags/claude-code) [View this post on daily.dev](https://daily.dev/posts/the-ai-native-sdlc-starts-with-your-infrastructure-2fviwpotm) ```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":"The AI-Native SDLC Starts With Your Infrastructure","url":"https://daily.dev/posts/the-ai-native-sdlc-starts-with-your-infrastructure-2fviwpotm","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/the-ai-native-sdlc-starts-with-your-infrastructure-2fviwpotm"},"datePublished":"2026-08-31T07:19:38.144Z","dateModified":"2026-09-13T21:00:25.503Z","description":"Anthropic's AI-Native SDLC playbook lays out six stages for coding-agent-driven development but never specifies what an agent's tests should actually run...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/41fb9dcb2dc2f1e44d1ec3805870544f?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/41fb9dcb2dc2f1e44d1ec3805870544f?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"MetalBear","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":"MetalBear","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/3165f5f876b147e2a160f4bf0fcaa13a","url":"https://daily.dev/sources/metalbear"},"commentCount":1,"discussionUrl":"https://daily.dev/posts/the-ai-native-sdlc-starts-with-your-infrastructure-2fviwpotm","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":1}],"keywords":"testing,kubernetes,ai-agents,claude-code","timeRequired":"PT6M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"MetalBear","item":"https://daily.dev/sources/metalbear"},{"@type":"ListItem","position":3,"name":"The AI-Native SDLC Starts With Your Infrastructure"}]} {"@context":"https://schema.org","@type":"WebPage","@id":"https://daily.dev/posts/the-ai-native-sdlc-starts-with-your-infrastructure-2fviwpotm","comment":[{"@type":"Comment","text":"I’ve had the fake billing service still return 200 after the real auth header changed. The agent saw green and called it done.","datePublished":"2026-09-01T03:09:50.140Z","url":"https://daily.dev/posts/2FViWpOtM#c-4Ymc8cnuf","author":{"@type":"Person","name":"Leonid Bugaev","url":"https://daily.dev/leonidbugaev","image":"https://lh3.googleusercontent.com/a/ACg8ocLhMgurwTmJElWaH9A8Ju8cHvZxgMCDt009jEYkmFCyUGAoaGSw=s96-c"}}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/the-ai-native-sdlc-starts-with-your-infrastructure-2fviwpotm#faq","mainEntity":[{"@type":"Question","name":"Why can't an AI coding agent be trusted to judge whether its tests passing against fake services actually means the code works in production?","acceptedAnswer":{"@type":"Answer","text":"An agent lacks the tacit knowledge a developer has about how stale a set of fakes has become, such as a mocked billing service still using an auth header that changed months ago, or a fake search endpoint that never paginates results the way the real one does. The agent only sees tests pass and reports the work finished, without knowing whether those fakes still resemble the real dependency. daily.dev surfaces practical takes like this for developers rethinking how agent-generated code gets verified."}},{"@type":"Question","name":"How does mirrord let AI coding agents test code against real staging services instead of local mocks?","acceptedAnswer":{"@type":"Answer","text":"mirrord proxies a locally running process's environment variables, secrets, and network traffic to and from a shared staging cluster, so the code behaves as if deployed there without actually being deployed. This lets an agent's outbound calls hit real upstream services and see real queue contents, rather than working against fakes on its own machine or relying on documentation. developers evaluating agent verification tooling can track approaches like mirrord's on daily.dev."}},{"@type":"Question","name":"How does mirrord handle multiple AI coding agent sessions running against the same shared staging cluster without interfering with each other?","acceptedAnswer":{"@type":"Answer","text":"The mirrord operator filters traffic by header so each agent session only receives its own requests, splits queues so each session gets a private slice of a shared topic, and branches databases so a session that writes does not disturb others. This lets one staging cluster serve many concurrent agent sessions instead of forcing each agent onto its own slow, expensive environment copy. teams scaling multi-agent coding workflows can follow infrastructure patterns like this via daily.dev."}}]} ```