--- title: From chaos to context: Building an AI dev workflow description: A GitLab engineer describes an evolving personal AI-assisted development workflow built around persistent, specific directives, git worktree-based parallel AI... canonical: https://daily.dev/posts/from-chaos-to-context-building-an-ai-dev-workflow-csabhxd1r twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: From chaos to context: Building an AI dev workflow | daily.dev og:description: A GitLab engineer describes an evolving personal AI-assisted development workflow built around persistent, specific directives, git worktree-based parallel AI... og:url: https://daily.dev/posts/from-chaos-to-context-building-an-ai-dev-workflow-csabhxd1r og:image: https://api.daily.dev/og/posts/CsabhXd1r.png og:image:alt: From chaos to context: Building an AI dev workflow 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. # From chaos to context: Building an AI dev workflow **[GitLab](https://daily.dev/sources/gitlab)** · 16 min read · 0 upvotes · 0 comments ## Summary A GitLab engineer describes an evolving personal AI-assisted development workflow built around persistent, specific directives, git worktree-based parallel AI sessions, and a custom semantic memory system called opencode-memory (built on MCP) that now integrates with GitLab's new Orbit SDLC knowledge graph. Key lessons include making directives explicit and specific rather than vague, using proactive context injection instead of manual recall (reaching ~91% effectiveness at surfacing relevant context automatically over 30 days), and treating human judgment and architectural oversight as the parts AI still cannot replace. The piece frames context management, not raw AI capability, as the real bottleneck in agentic coding workflows. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How can I give an AI coding assistant persistent memory across sessions instead of re-explaining context every time? Build a semantic memory system with hybrid search combining keyword and vector recall, session coordination through claim/release primitives, and automatic boot context that surfaces directives. One implementation, opencode-memory, grew to a knowledge graph of over 760,000 indexed code entities and 15,000+ memories, and later integrated with GitLab's Orbit SDLC graph for cross-project context. _daily.dev surfaces workflow write-ups like this for teams wiring persistent memory into their coding agents._ ### How effective is proactive context injection at reducing manual recall prompts for AI coding agents? Over a 30-day trial, a system that automatically searched for relevant memories before each interaction achieved around 91% effectiveness at surfacing relevant context without prompting. Sessions using proactive injection needed zero explicit recall calls on average, compared to 17 explicit recall calls per session without it. _Developers tuning agent context strategies compare real numbers like these on daily.dev before rebuilding their own setup._ ### Why did GitLab replace its CREDIT values with new operating principles for AI use? GitLab retired CREDIT entirely and replaced it with new operating principles built for the agentic era, announced as part of GitLab Act 2. The change happened alongside internal guidance work on how CREDIT applied to AI use, illustrating how quickly organizational AI policy can shift underneath in-progress work. _Teams tracking how vendors adapt policy for agentic AI can follow shifts like this one on daily.dev._ ## Similar posts on daily.dev - [Bridging the Gap - How We Made AI Agents 10x Developers in Our Organization](https://daily.dev/posts/bridging-the-gap---how-we-made-ai-agents-10x-developers-in-our-organization-9xldqiaaq) · DEV · 4 upvotes · 0 comments - [My Thoughts on AI, Part 2: Agent Setup, Workflow, and Tools](https://daily.dev/posts/my-thoughts-on-ai-part-2-agent-setup-workflow-and-tools-bwcpsofep) · Mark Erikson · 0 upvotes · 0 comments - [I struggled to code with AI until I learned this workflow](https://daily.dev/posts/i-struggled-to-code-with-ai-until-i-learned-this-workflow-habxlgdyu) · System Design Newsletter · 90 upvotes · 3 comments - [Beyond Code Generation: AI for the Full Data Science Workflow](https://daily.dev/posts/beyond-code-generation-ai-for-the-full-data-science-workflow-hqrcj0p03) · Towards Data Science · 1 upvotes · 0 comments --- Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#ai-coding](https://daily.dev/tags/ai-coding), [#mcp](https://daily.dev/tags/mcp), [#gitlab](https://daily.dev/tags/gitlab), [#opencode](https://daily.dev/tags/opencode) [View this post on daily.dev](https://daily.dev/posts/from-chaos-to-context-building-an-ai-dev-workflow-csabhxd1r) ```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":"From chaos to context: Building an AI dev workflow","url":"https://daily.dev/posts/from-chaos-to-context-building-an-ai-dev-workflow-csabhxd1r","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/from-chaos-to-context-building-an-ai-dev-workflow-csabhxd1r"},"datePublished":"2026-08-19T18:51:31.088Z","dateModified":"2026-09-14T07:42:50.071Z","description":"A GitLab engineer describes an evolving personal AI-assisted development workflow built around persistent, specific directives, git worktree-based parallel AI...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/a474c7e279317af818b54220398fc4e7?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/a474c7e279317af818b54220398fc4e7?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"GitLab","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":"GitLab","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/3720ce340ca4435daef529abbf361b69","url":"https://daily.dev/sources/gitlab"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/from-chaos-to-context-building-an-ai-dev-workflow-csabhxd1r","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai-agents,ai-coding,mcp,gitlab,opencode","timeRequired":"PT16M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"GitLab","item":"https://daily.dev/sources/gitlab"},{"@type":"ListItem","position":3,"name":"From chaos to context: Building an AI dev workflow"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/from-chaos-to-context-building-an-ai-dev-workflow-csabhxd1r#faq","mainEntity":[{"@type":"Question","name":"How can I give an AI coding assistant persistent memory across sessions instead of re-explaining context every time?","acceptedAnswer":{"@type":"Answer","text":"Build a semantic memory system with hybrid search combining keyword and vector recall, session coordination through claim/release primitives, and automatic boot context that surfaces directives. One implementation, opencode-memory, grew to a knowledge graph of over 760,000 indexed code entities and 15,000+ memories, and later integrated with GitLab's Orbit SDLC graph for cross-project context. daily.dev surfaces workflow write-ups like this for teams wiring persistent memory into their coding agents."}},{"@type":"Question","name":"How effective is proactive context injection at reducing manual recall prompts for AI coding agents?","acceptedAnswer":{"@type":"Answer","text":"Over a 30-day trial, a system that automatically searched for relevant memories before each interaction achieved around 91% effectiveness at surfacing relevant context without prompting. Sessions using proactive injection needed zero explicit recall calls on average, compared to 17 explicit recall calls per session without it. Developers tuning agent context strategies compare real numbers like these on daily.dev before rebuilding their own setup."}},{"@type":"Question","name":"Why did GitLab replace its CREDIT values with new operating principles for AI use?","acceptedAnswer":{"@type":"Answer","text":"GitLab retired CREDIT entirely and replaced it with new operating principles built for the agentic era, announced as part of GitLab Act 2. The change happened alongside internal guidance work on how CREDIT applied to AI use, illustrating how quickly organizational AI policy can shift underneath in-progress work. Teams tracking how vendors adapt policy for agentic AI can follow shifts like this one on daily.dev."}}]} ```