--- title: Your SDLC is your context engineering | daily.dev description: A founding engineer at Vendict describes how the team rebuilt its SDLC specifically to support AI agents as first-class contributors. The core thesis:... canonical: https://daily.dev/posts/your-sdlc-is-your-context-engineering-w6tqmjmyc twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Your SDLC is your context engineering | daily.dev og:description: A founding engineer at Vendict describes how the team rebuilt its SDLC specifically to support AI agents as first-class contributors. The core thesis:... og:url: https://daily.dev/posts/your-sdlc-is-your-context-engineering-w6tqmjmyc og:image: https://api.daily.dev/og/posts/w6tqMjMyC.png og:image:alt: Your SDLC is your context engineering 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. # Your SDLC is your context engineering **[LeadDev](https://daily.dev/sources/leaddev)** · 19 min read · 15 upvotes · 1 comments ## Summary A founding engineer at Vendict describes how the team rebuilt its SDLC specifically to support AI agents as first-class contributors. The core thesis: file-level context (AGENTS.md) is insufficient — the real context for agents is the entire development lifecycle, including how work is specified, reviewed, tested, and shipped. Key design choices include: separating builder and reviewer roles at the task level (not just PR level) using subagents with isolated contexts; excluding test files from PR size limits while using a separate test-design subagent; treating spec.md and plan.md as distinct artifacts with separate review surfaces; and maintaining a single standards layer in docs/standards/ consumed by all tools. After a month of real use, the team found that nearly every critical issue traced back to planning and specification quality, not code generation. The next bottleneck identified is code review throughput as agents produce more focused PRs faster than humans can review them. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### Why should test files be excluded from pull request size limits when using AI code review tools? Excluding tests from PR size caps prevents agents from writing fewer tests or splitting tests into separate PRs just to stay under the limit, both of which undermine thorough test coverage. Defect detection in review drops from 87% on small diffs to 28% on diffs over a thousand lines, so size limits matter, but test design is instead handled by a separate subagent with no line limit that plans edge cases and failure modes independently of the implementation agent's reasoning. _Teams weighing PR size policy for AI-heavy review can compare real-world setups like this on daily.dev._ ### Why separate the builder and reviewer roles at the task level instead of just at the pull request level when using AI coding agents? A single coding session today can span implementation, tests, verification, and lint, all within one agent's context, so separating builder and reviewer only at the PR level is insufficient since the same reasoning that decided an implementation was correct will also rationalize away test failures. Enforcing separation at the task level uses four distinct subagents (test designer, failure analyzer, verifier, linter), each in its own context, with unclear failures escalated to a human rather than resolved automatically. _Engineers designing agent review workflows can track patterns like task-level separation on daily.dev._ ### What is the most common root cause of critical issues in codebases developed heavily with AI coding agents? Almost every critical issue traces back to the specification and insufficiently thoughtful planning rather than the code itself. This finding emerged after a month of running a redesigned software development lifecycle at a startup, shifting the emphasis of quality efforts toward how clearly humans define the work before agents write any code. _Anyone debugging where AI-assisted development goes wrong can follow planning-versus-code debates on daily.dev._ ## Community discussion Top comments from developers on daily.dev. **@agustinbarrientos** · 0 upvotes > So the real prompt has been hiding in Jira this whole time ## Similar posts on daily.dev - [The AI SDLC transformation playbook](https://daily.dev/posts/the-ai-sdlc-transformation-playbook-hykxezayi) · Atlassian · 1 upvotes · 0 comments - [Your agent context needs a development lifecycle](https://daily.dev/posts/your-agent-context-needs-a-development-lifecycle-vacgztkaf) · The New Stack · 0 upvotes · 1 comments - [The New Software Lifecycle](https://daily.dev/posts/the-new-software-lifecycle-hwvvtnaz2) · Addy Osmani · 12 upvotes · 1 comments - [The Context Tax of Agentic Development](https://daily.dev/posts/the-context-tax-of-agentic-development-spfv0nbhr) · Expedia Group Tech · 0 upvotes · 0 comments - [ADLC: The Lifecycle Taking Shape](https://daily.dev/posts/adlc-the-lifecycle-taking-shape-napaas7zg) · Medium · 0 upvotes · 0 comments --- Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#code-review](https://daily.dev/tags/code-review), [#context-engineering](https://daily.dev/tags/context-engineering) [View this post on daily.dev](https://daily.dev/posts/your-sdlc-is-your-context-engineering-w6tqmjmyc) ```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":"Your SDLC is your context engineering","url":"https://daily.dev/posts/your-sdlc-is-your-context-engineering-w6tqmjmyc","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/your-sdlc-is-your-context-engineering-w6tqmjmyc"},"datePublished":"2026-08-10T12:25:43.202Z","dateModified":"2026-09-13T20:57:23.069Z","description":"A founding engineer at Vendict describes how the team rebuilt its SDLC specifically to support AI agents as first-class contributors. The core thesis:...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/cf528fd511ee329ac639c1b6e63252ba?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/cf528fd511ee329ac639c1b6e63252ba?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"LeadDev","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":"LeadDev","logo":"https://media.daily.dev/image/upload/s--u_Q34nzh--/f_auto,q_auto/v1774514111/logos/leaddev?_a=BAMAMiWQ0","url":"https://daily.dev/sources/leaddev"},"commentCount":1,"discussionUrl":"https://daily.dev/posts/your-sdlc-is-your-context-engineering-w6tqmjmyc","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":15},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":1}],"keywords":"ai-agents,code-review,context-engineering","timeRequired":"PT19M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"LeadDev","item":"https://daily.dev/sources/leaddev"},{"@type":"ListItem","position":3,"name":"Your SDLC is your context engineering"}]} {"@context":"https://schema.org","@type":"WebPage","@id":"https://daily.dev/posts/your-sdlc-is-your-context-engineering-w6tqmjmyc","comment":[{"@type":"Comment","text":"So the real prompt has been hiding in Jira this whole time","datePublished":"2026-08-21T06:25:38.461Z","url":"https://daily.dev/posts/w6tqMjMyC#c-hMom1MxTq","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/your-sdlc-is-your-context-engineering-w6tqmjmyc#faq","mainEntity":[{"@type":"Question","name":"Why should test files be excluded from pull request size limits when using AI code review tools?","acceptedAnswer":{"@type":"Answer","text":"Excluding tests from PR size caps prevents agents from writing fewer tests or splitting tests into separate PRs just to stay under the limit, both of which undermine thorough test coverage. Defect detection in review drops from 87% on small diffs to 28% on diffs over a thousand lines, so size limits matter, but test design is instead handled by a separate subagent with no line limit that plans edge cases and failure modes independently of the implementation agent's reasoning. Teams weighing PR size policy for AI-heavy review can compare real-world setups like this on daily.dev."}},{"@type":"Question","name":"Why separate the builder and reviewer roles at the task level instead of just at the pull request level when using AI coding agents?","acceptedAnswer":{"@type":"Answer","text":"A single coding session today can span implementation, tests, verification, and lint, all within one agent's context, so separating builder and reviewer only at the PR level is insufficient since the same reasoning that decided an implementation was correct will also rationalize away test failures. Enforcing separation at the task level uses four distinct subagents (test designer, failure analyzer, verifier, linter), each in its own context, with unclear failures escalated to a human rather than resolved automatically. Engineers designing agent review workflows can track patterns like task-level separation on daily.dev."}},{"@type":"Question","name":"What is the most common root cause of critical issues in codebases developed heavily with AI coding agents?","acceptedAnswer":{"@type":"Answer","text":"Almost every critical issue traces back to the specification and insufficiently thoughtful planning rather than the code itself. This finding emerged after a month of running a redesigned software development lifecycle at a startup, shifting the emphasis of quality efforts toward how clearly humans define the work before agents write any code. Anyone debugging where AI-assisted development goes wrong can follow planning-versus-code debates on daily.dev."}}]} ```