--- title: Three Levels for AI Agents: The Repo, the Company, and... description: Proposes organizing AI agent guidance into three layers: repository-level rules (architecture, code style, exceptions specific to an app), company-level shared... canonical: https://daily.dev/posts/three-levels-for-ai-agents-the-repo-the-company-and-the-developer-wbcjazeiw twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Three Levels for AI Agents: The Repo, the Company, and the Developer | daily.dev og:description: Proposes organizing AI agent guidance into three layers: repository-level rules (architecture, code style, exceptions specific to an app), company-level shared... og:url: https://daily.dev/posts/three-levels-for-ai-agents-the-repo-the-company-and-the-developer-wbcjazeiw og:image: https://api.daily.dev/og/posts/WbCJAzeiw.png og:image:alt: Three Levels for AI Agents: The Repo, the Company, and the Developer 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. # Three Levels for AI Agents: The Repo, the Company, and the Developer **[Medium](https://daily.dev/sources/medium_js)** · 8 min read · 13 upvotes · 3 comments ## Summary Proposes organizing AI agent guidance into three layers: repository-level rules (architecture, code style, exceptions specific to an app), company-level shared workflows (ticket handling, PR process), and a personal developer layer capturing individual expertise and review heuristics. Argues that standardizing collaboration should not flatten developer judgment, and that code review remains the place where diverse experience becomes shared knowledge. References Bacchelli and Bird's study on code review outcomes, Google's review guidelines distinguishing mandatory vs. educational comments, and Anthropic's Agent Skills concept for on-demand context loading. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How should teams organize AI coding agent instructions across different projects and developers? Instructions can be split into three layers: repository-level rules covering architecture and code style specific to an app, a company-wide shared workflow for ticket handling and PR process, and a personal developer layer holding individual review heuristics and experience. Contradictions between layers should be surfaced and resolved rather than letting one silently override another. _Teams standardizing AI agent workflows while preserving individual judgment can track approaches like this on daily.dev._ ### What does Anthropic's Agent Skills concept do for managing context in AI agents? Agent Skills uses on-demand loading: short descriptions make available skills discoverable first, and detailed instructions or resources are only read in when actually needed, rather than loading everything upfront. This reduces the amount of context an agent processes for a given task while keeping deeper documentation accessible when required. _Developers tuning agent context budgets can follow practical approaches like this via daily.dev._ ## Community discussion Top comments from developers on daily.dev. **@skepagent** · 1 upvotes > The repo layer is the one I'd double down on, but with one caveat: agent instructions rot. A rule that was right in March can be actively wrong by September, and nobody re-reads AGENTS.md on a schedule. Worth a periodic 'does this still apply' prune — stale guidance is worse than none. **@alexx\_dev** · 0 upvotes > do you tie the repo-level guidance lifecycle to code releases or keep a separate audit schedule, and how do you avoid stale rules breaking downstream pipelines? ## Similar posts on daily.dev - [Why AI-driven development still demands human oversight](https://daily.dev/posts/why-ai-driven-development-still-demands-human-oversight-tvvjeefam) · SD Times · 1 upvotes · 0 comments - [Agentic Code Review](https://daily.dev/posts/agentic-code-review-0ue8fcsa1) · Addy Osmani · 146 upvotes · 14 comments - [Medium](https://daily.dev/posts/medium-7efociuwx) · Medium · 0 upvotes · 0 comments - [Level 3 AI Collaboration: The Sweet Spot for Developers](https://daily.dev/posts/level-3-ai-collaboration-the-sweet-spot-for-developers-ekdkuydju) · The New Stack · 1 upvotes · 0 comments --- Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#architecture](https://daily.dev/tags/architecture), [#code-review](https://daily.dev/tags/code-review), [#company-culture](https://daily.dev/tags/company-culture) [View this post on daily.dev](https://daily.dev/posts/three-levels-for-ai-agents-the-repo-the-company-and-the-developer-wbcjazeiw) ```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":"Three Levels for AI Agents: The Repo, the Company, and the Developer","url":"https://daily.dev/posts/three-levels-for-ai-agents-the-repo-the-company-and-the-developer-wbcjazeiw","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/three-levels-for-ai-agents-the-repo-the-company-and-the-developer-wbcjazeiw"},"datePublished":"2026-10-16T18:27:49.037Z","dateModified":"2026-10-16T19:00:09.462Z","description":"Proposes organizing AI agent guidance into three layers: repository-level rules (architecture, code style, exceptions specific to an app), company-level shared...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/9e2a7810e67de52c9c874fef78931106?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/9e2a7810e67de52c9c874fef78931106?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Medium","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":"Medium","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/medium","url":"https://daily.dev/sources/medium_js"},"commentCount":3,"discussionUrl":"https://daily.dev/posts/three-levels-for-ai-agents-the-repo-the-company-and-the-developer-wbcjazeiw","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":13},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":3}],"keywords":"ai-agents,architecture,code-review,company-culture","timeRequired":"PT8M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Medium","item":"https://daily.dev/sources/medium_js"},{"@type":"ListItem","position":3,"name":"Three Levels for AI Agents: The Repo, the Company, and the Developer"}]} {"@context":"https://schema.org","@type":"WebPage","@id":"https://daily.dev/posts/three-levels-for-ai-agents-the-repo-the-company-and-the-developer-wbcjazeiw","comment":[{"@type":"Comment","text":"The repo layer is the one I’d double down on, but with one caveat: agent instructions rot. A rule that was right in March can be actively wrong by September, and nobody re-reads AGENTS.md on a schedule. Worth a periodic ‘does this still apply’ prune — stale guidance is worse than none.","datePublished":"2026-10-17T02:56:03.293Z","url":"https://daily.dev/posts/WbCJAzeiw#c-aOguOSqIP","author":{"@type":"Person","name":"kris gu","url":"https://daily.dev/skepagent","image":"https://lh3.googleusercontent.com/a/ACg8ocJnnwFKNx7PJNhb7W6mFWr2k2SV_VHW_9ainY71DX_dv79frQ=s96-c"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1}},{"@type":"Comment","text":"do you tie the repo-level guidance lifecycle to code releases or keep a separate audit schedule, and how do you avoid stale rules breaking downstream pipelines?","datePublished":"2026-10-18T02:00:54.747Z","url":"https://daily.dev/posts/WbCJAzeiw#c-nTZkNngok","author":{"@type":"Person","name":"alex","url":"https://daily.dev/alexx_dev"}}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/three-levels-for-ai-agents-the-repo-the-company-and-the-developer-wbcjazeiw#faq","mainEntity":[{"@type":"Question","name":"How should teams organize AI coding agent instructions across different projects and developers?","acceptedAnswer":{"@type":"Answer","text":"Instructions can be split into three layers: repository-level rules covering architecture and code style specific to an app, a company-wide shared workflow for ticket handling and PR process, and a personal developer layer holding individual review heuristics and experience. Contradictions between layers should be surfaced and resolved rather than letting one silently override another. Teams standardizing AI agent workflows while preserving individual judgment can track approaches like this on daily.dev."}},{"@type":"Question","name":"What does Anthropic's Agent Skills concept do for managing context in AI agents?","acceptedAnswer":{"@type":"Answer","text":"Agent Skills uses on-demand loading: short descriptions make available skills discoverable first, and detailed instructions or resources are only read in when actually needed, rather than loading everything upfront. This reduces the amount of context an agent processes for a given task while keeping deeper documentation accessible when required. Developers tuning agent context budgets can follow practical approaches like this via daily.dev."}}]} ```