--- title: The harness is all you need (mostly) | daily.dev description: A practical workflow for using GitHub Copilot effectively without chasing every new AI tool. The approach covers eight steps: picking a tool, enabling... canonical: https://daily.dev/posts/the-harness-is-all-you-need-mostly--vowadggsu twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: The harness is all you need (mostly) | daily.dev og:description: A practical workflow for using GitHub Copilot effectively without chasing every new AI tool. The approach covers eight steps: picking a tool, enabling... og:url: https://daily.dev/posts/the-harness-is-all-you-need-mostly--vowadggsu og:image: https://api.daily.dev/og/posts/vOWadggsu.png og:image:alt: The harness is all you need (mostly) 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 harness is all you need (mostly) **[GitHub Blog](https://daily.dev/sources/ghblog)** · 14 min read · 2 upvotes · 0 comments ## Summary A practical workflow for using GitHub Copilot effectively without chasing every new AI tool. The approach covers eight steps: picking a tool, enabling YOLO/allow-all mode for agent autonomy, prototyping visually before coding, planning methodically using plan mode to surface edge cases, implementing with Autopilot's built-in loop, iterating through human review, performing a Rubber Duck cross-model review, and committing the result. The core argument is that mastering the existing harness — the fundamental interaction model — yields more productivity gains than constantly adopting new tools, prompts, or configurations. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is Rubber Duck review in GitHub Copilot? Rubber Duck review is a GitHub Copilot feature where a different AI model family reviews completed work, such as Sonnet reviewing output generated by GPT 5.6 Terra. Because different models are trained on different data, they have different blind spots, so this second-model review surfaces issues a single model would miss. It can be applied to prototypes, plans, or final implementations, and can be combined with Autopilot to loop reviews until only diminishing-return items remain. _See how developers combine multi-model review loops into their AI coding workflow on daily.dev._ ### What is YOLO mode in an AI coding agent and why use it? YOLO mode, also called 'Allow All', lets a coding agent execute commands without asking for approval each time, which is necessary for real productivity gains since constantly approving actions trains you to stop reading them anyway. It should not be run on a local machine, especially at work, since agents then have unchecked autonomy over private organizational data and mistakes can be costly. _Developers weighing agent autonomy against safety can track workflow practices like this on daily.dev._ ### How does GitHub Copilot's Autopilot mode work with subagents? Autopilot is a built-in loop that forces a model to keep working until it has completed every item in an agreed plan. GitHub Copilot automatically orchestrates this by delegating file-reading tasks to an 'Explore' subagent using a small model, and more complex actions to a 'General Purpose' subagent using a larger model, without requiring any manual configuration for multimodel orchestration. _Anyone tuning agent orchestration and subagent behavior can follow practical patterns like this on daily.dev._ ## Similar posts on daily.dev - [Reverse Engineering GitHub Copilot: What Every AI Product Manager Can Learn](https://daily.dev/posts/reverse-engineering-github-copilot-what-every-ai-product-manager-can-learn-l0nf4chfi) · Medium · 0 upvotes · 0 comments --- Tags: [#github](https://daily.dev/tags/github), [#ai-agents](https://daily.dev/tags/ai-agents), [#ai-coding](https://daily.dev/tags/ai-coding), [#web-components](https://daily.dev/tags/web-components) [View this post on daily.dev](https://daily.dev/posts/the-harness-is-all-you-need-mostly--vowadggsu) ```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 harness is all you need (mostly)","url":"https://daily.dev/posts/the-harness-is-all-you-need-mostly--vowadggsu","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/the-harness-is-all-you-need-mostly--vowadggsu"},"datePublished":"2026-07-27T18:03:16.685Z","dateModified":"2026-09-13T20:06:39.772Z","description":"A practical workflow for using GitHub Copilot effectively without chasing every new AI tool. The approach covers eight steps: picking a tool, enabling...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/6ba2f60e52e8020068ab76bbc8b720e3?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/6ba2f60e52e8020068ab76bbc8b720e3?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"GitHub Blog","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":"GitHub Blog","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/106cf162b88840808484d4b5429b59b1","url":"https://daily.dev/sources/ghblog"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/the-harness-is-all-you-need-mostly--vowadggsu","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":2},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"github,ai-agents,ai-coding,web-components","timeRequired":"PT14M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"GitHub Blog","item":"https://daily.dev/sources/ghblog"},{"@type":"ListItem","position":3,"name":"The harness is all you need (mostly)"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/the-harness-is-all-you-need-mostly--vowadggsu#faq","mainEntity":[{"@type":"Question","name":"What is Rubber Duck review in GitHub Copilot?","acceptedAnswer":{"@type":"Answer","text":"Rubber Duck review is a GitHub Copilot feature where a different AI model family reviews completed work, such as Sonnet reviewing output generated by GPT 5.6 Terra. Because different models are trained on different data, they have different blind spots, so this second-model review surfaces issues a single model would miss. It can be applied to prototypes, plans, or final implementations, and can be combined with Autopilot to loop reviews until only diminishing-return items remain. See how developers combine multi-model review loops into their AI coding workflow on daily.dev."}},{"@type":"Question","name":"What is YOLO mode in an AI coding agent and why use it?","acceptedAnswer":{"@type":"Answer","text":"YOLO mode, also called 'Allow All', lets a coding agent execute commands without asking for approval each time, which is necessary for real productivity gains since constantly approving actions trains you to stop reading them anyway. It should not be run on a local machine, especially at work, since agents then have unchecked autonomy over private organizational data and mistakes can be costly. Developers weighing agent autonomy against safety can track workflow practices like this on daily.dev."}},{"@type":"Question","name":"How does GitHub Copilot's Autopilot mode work with subagents?","acceptedAnswer":{"@type":"Answer","text":"Autopilot is a built-in loop that forces a model to keep working until it has completed every item in an agreed plan. GitHub Copilot automatically orchestrates this by delegating file-reading tasks to an 'Explore' subagent using a small model, and more complex actions to a 'General Purpose' subagent using a larger model, without requiring any manual configuration for multimodel orchestration. Anyone tuning agent orchestration and subagent behavior can follow practical patterns like this on daily.dev."}}]} ```