Use Claude, Codex and Copilot Together on GitHub  [[3. Resources/Readwise/Full Document Contents/Articles/Use Claude, Codex and Copilot Together on GitHub|Full Document Contents]] ## Document Notes > [!note]+ Document Notes > tags: [github-copilot, ai-agents, development-workflow] > author: GitHub > source: YouTube > reference: https://www.youtube.com/watch?v=GuTQDXKwdJQ > date created: 2024-12-19 19:30 > > # 🏷️ Title: Use Claude, Codex and Copilot together on GitHub > <!-- deep-read:sections start --> ## 🧰 Action Inbox - [H1] Set up agent orchestration workflow — explore GitHub agents tab #ai-tools - [H2] Test parallel agent assignments — compare Claude vs Codex vs Copilot performance #workflow - [H3] Configure mobile agent management — enable remote task assignment #productivity ## 🧭 Table of Contents - [📝 Summary](#-summary) - [📖 Main Content](#-main-content) - [🗒️ Notes](#️-notes) - [📚 References](#-references) - [🗣️ Paraphrase](#️-paraphrase) - [📰 Newsletter](#-newsletter) - [✅ Action Items](#-action-items) - [🧪 Hypotheses](#-hypotheses) - [📊 Methodology](#-methodology) - [❓ Questions](#-questions) - [🔮 Future Research](#-future-research) - [💡 Implications](#-implications) - [🧠 Learning Styles Assessment](#-learning-styles-assessment) - [📊 Accessibility Assessment](#-accessibility-assessment) - [🌍 DEIB Assessment](#-deib-assessment) - [✅ Logical Validity](#-logical-validity) - [🛡️ DARVO Analysis](#-darvo-analysis) - [🎯 Conclusion](#-conclusion) ### 🔗 References - [GitHub Copilot](https://github.com/features/copilot) - [Claude by Anthropic](https://claude.ai) ## 📝 Summary - GitHub has introduced Agent HQ allowing developers to orchestrate multiple AI coding assistants simultaneously - Developers can assign tasks to Claude, Codex, or Copilot in parallel rather than sequentially - The system integrates across GitHub web, mobile, and VS Code IDE - Tasks can be assigned directly from issues with branch selection - This represents a shift from single-agent workflows to multi-agent orchestration ## 📖 Main Content GitHub's new Agent HQ functionality enables developers to manage multiple AI coding assistants from a centralized location. Instead of working with one AI assistant at a time, developers can now assign different coding tasks to Claude (by Anthropic), GitHub Copilot, or Codex simultaneously. This parallel processing approach aims to improve development velocity and leverage the unique strengths of different AI models. The system integrates deeply into existing GitHub workflows, allowing task assignment directly from issues and providing mobile access for on-the-go management. VS Code integration ensures developers can orchestrate agent sessions without leaving their primary development environment. ## 🗒️ Notes #### ⏱️ [00:00](https://youtu.be/GuTQDXKwdJQ?t=0) — [H1] Introduction to parallel agent workflow - Key idea: GitHub Copilot enables shifting from sequential to simultaneous development tasks - Evidence: "you can shift your development tasks from sequential to simultaneous with the agent of your choice" #### ⏱️ [00:08](https://youtu.be/GuTQDXKwdJQ?t=8) — [H2] Agent HQ centralized management - Key idea: Agents tab provides unified control for all cloud agent sessions - Evidence: "The agents tab on GitHub is where you can manage all your cloud agent session" #### ⏱️ [00:15](https://youtu.be/GuTQDXKwdJQ?t=15) — [H2] Multi-agent task assignment - Key idea: Tasks can be assigned to different AI models from one interface - Evidence: "you can assign ad hoc tasks to copilot to claw byanthropic to codeex or to a custom agent" #### ⏱️ [00:24](https://youtu.be/GuTQDXKwdJQ?t=24) — [H3] Centralized monitoring capabilities - Key idea: All agent tasks can be monitored from single GitHub location - Evidence: "monitor our tasks all from one centralized location on GitHub" #### ⏱️ [00:30](https://youtu.be/GuTQDXKwdJQ?t=30) — [H2] Issue-based agent assignment - Key idea: GitHub issues can be directly assigned to specific AI agents - Evidence: "you can assign issues directly to an agent like copilot codeex or claude" #### ⏱️ [00:37](https://youtu.be/GuTQDXKwdJQ?t=37) — [H3] Branch and agent selection workflow - Key idea: Developers choose both the base branch and preferred agent - Evidence: "Choose the branch you wish to use as the base and again the agent you want" #### ⏱️ [00:43](https://youtu.be/GuTQDXKwdJQ?t=43) — [H2] Mobile agent management - Key idea: Agent orchestration extends to mobile GitHub app - Evidence: "We can manage agents and even assign tasks right from GitHub mobile" #### ⏱️ [00:52](https://youtu.be/GuTQDXKwdJQ?t=52) — [H1] VS Code IDE integration - Key idea: Developers can orchestrate agent sessions without leaving their IDE - Evidence: "In VS Code, you can orchestrate all your agent sessions local and cloud from your IDE" #### ⏱️ [01:02](https://youtu.be/GuTQDXKwdJQ?t=62) — [H3] Custom agent and model flexibility - Key idea: System supports custom agents and specified models beyond default options - Evidence: "assign to claude or codeex or to copilot using a custom agent or specified model" #### ⏱️ [01:10](https://youtu.be/GuTQDXKwdJQ?t=70) — [H1] Workflow integration summary - Key idea: Agent management integrates seamlessly into existing development workflows - Evidence: "assign and manage parallel coding agent sessions all within your existing workflow" ## 📚 References - [GitHub Copilot](https://github.com/features/copilot) - [Claude by Anthropic](https://claude.ai) - [GitHub Mobile](https://github.com/mobile) - [VS Code](https://code.visualstudio.com/) ## 🗣️ Paraphrase - GitHub has created a unified interface called Agent HQ where developers can manage multiple AI coding assistants from one location - Instead of using AI tools one at a time, developers can now run multiple agents simultaneously on different tasks - The system works across all GitHub platforms including web, mobile, and VS Code, maintaining consistency in existing workflows ## 📰 Newsletter **Multi-Agent AI Development Comes to GitHub** GitHub's new Agent HQ transforms how developers work with AI assistants. Instead of switching between Claude, Copilot, and Codex, you can now orchestrate all three simultaneously from a single interface. **Key Benefits:** - Assign different tasks to different AI models based on their strengths - Monitor all agent sessions from one centralized GitHub location - Work seamlessly across web, mobile, and VS Code environments - Integrate agent assignments directly into your issue workflow **Getting Started:** - Access the new Agents tab in your GitHub repository - Try assigning issues to different agents to compare results - Experiment with parallel task assignment for complex features This represents a significant shift toward multi-agent development workflows that could dramatically improve coding velocity and quality. Resources: [GitHub Copilot](https://github.com/features/copilot), [VS Code](https://code.visualstudio.com/), [GitHub Mobile](https://github.com/mobile) ## ✅ Action Items #### DevRel - Developer advocacy: Create demos showing multi-agent workflow benefits, track adoption metrics - Education: Develop workshops on agent orchestration best practices, comparative model strengths - Developer experience: Document agent selection guidelines, create troubleshooting resources - Developer marketing: Campaign highlighting productivity gains from parallel AI assistance - Community management: Gather feedback on agent preferences, moderate discussions on AI tool effectiveness #### OSPO - Policy & Governance: Review AI agent usage policies, establish guidelines for external model integration - License Compliance: Audit licensing terms for Claude, Codex usage in enterprise environments - Security Management: Assess data exposure risks across multiple AI models, review API security - Software Inventory: Track AI agent dependencies, document model version compatibility - Developer Support: Train teams on multi-agent workflows, provide troubleshooting guidance - Community Engagement: Evaluate impact on open source contribution patterns - Strategic Alignment: Define KPIs for multi-agent adoption, measure productivity improvements #### Community - Maintainers: Develop templates for agent-assisted code review, establish quality standards - Contributors: Create onboarding guides for AI-assisted contributions, define acceptable use policies - Sponsors: Assess cost implications of multi-agent usage, budget for increased AI compute - Ecosystem: Coordinate standards for agent interoperability, share best practices #### WordPress - Core: Experiment with agent-assisted patch development, evaluate code quality impacts - Core Performance: Use agents for optimization testing, automated performance analysis - Core AI: Integrate agent capabilities into WordPress development workflows - Documentation: Generate agent-assisted documentation updates, improve technical writing - CLI: Explore agent integration for command development, automated testing - Plugins: Develop agent-assisted plugin review processes, security scanning - Meta: Assess agent capabilities for .org platform maintenance ## 🧪 Hypotheses - Multi-agent workflows will increase development velocity by 30-50% compared to single-agent approaches - Different AI models will show distinct strengths for specific coding tasks (Claude for documentation, Copilot for code completion) - Mobile agent management will increase developer productivity during non-desktop hours - Parallel agent assignment will reduce context switching overhead in complex projects ### 📊 Methodology - Study design: Compare development velocity metrics before/after multi-agent adoption across representative projects - Data sources: GitHub commit frequency, issue resolution time, code review cycles, developer surveys - Experiments: A/B test single-agent vs multi-agent workflows on similar feature development tasks ### ❓ Questions - How will developers decide which agent to assign to specific tasks? - What happens when multiple agents produce conflicting solutions for the same problem? - How does cost scale with parallel agent usage compared to sequential workflows? ### 🔮 Future Research - Agent specialization patterns and optimal task allocation strategies - Integration possibilities with other development platforms beyond GitHub - Long-term impact on developer skill development and learning ### 💡 Implications - Practice: Fundamental shift in how developers approach AI-assisted coding workflows - Policy: Need for clear guidelines on appropriate agent selection and usage - Security: Expanded attack surface with multiple AI model integrations - Education: Training requirements for effective multi-agent orchestration - Funding: Increased costs for AI compute across multiple models simultaneously ### 🧠 Learning Styles Assessment Reading level estimate: 8th grade (accessible technical content) #### 👀 Visual learners - Present: Interface screenshots of agents tab, workflow diagrams implied - Improve: Architecture diagrams showing agent interaction, visual comparison charts #### 🔊 Aural learners - Present: Video demonstration with narration - Improve: Podcast-style explanation of benefits, audio tutorials #### 🗣️ Verbal learners - Present: Clear explanations of agent capabilities, defined terminology - Improve: Analogies comparing agents to different specialist roles #### 🧩 Kinesthetic learners - Present: Step-by-step workflow descriptions - Improve: Interactive sandbox for testing agent assignments, hands-on labs #### 🧠 Logical learners - Present: Sequential workflow explanations, cause-effect relationships - Improve: Decision trees for agent selection, performance comparison tables #### 🧑‍🤝‍🧑 Social learners - Present: Collaborative workflow implications - Improve: Team coordination guides, peer review integration #### 🧘‍♀️ Solitary learners - Present: Individual productivity focus - Improve: Self-paced learning modules, personal workflow optimization #### 🌿 Naturalistic learners - Present: Workflow ecosystem thinking - Improve: Development environment integration analogies ### 📊 Accessibility Assessment #### Screen reader semantics - Present: Video has spoken content describing interface elements - Improve: Detailed alt text for interface screenshots, structured heading navigation #### Keyboard navigation - Present: GitHub interface generally keyboard accessible - Improve: Ensure agent selection dropdowns are keyboard navigable #### Contrast and readability - Present: GitHub's interface meets basic contrast requirements - Improve: Verify agent status indicators have sufficient color contrast #### Text resizing and reflow - Present: GitHub interface responsive to zoom - Improve: Ensure agent management panels reflow properly at 200% zoom #### Media and captions - Present: Video content available - Improve: Add captions for accessibility, provide transcript #### Forms and interactions - Present: Agent selection forms with labels - Improve: Error handling for failed agent assignments, clear feedback ### 🌍 DEIB Assessment #### Cultural breadth - Present: Global platform accessibility - Improve: Examples from diverse development contexts, multilingual agent support #### Inclusive language - Present: Technical terminology clearly defined - Improve: Avoid assumptions about developer experience levels #### Neurodiversity support - Present: Structured workflow descriptions - Improve: Clear visual organization, consistent interface patterns #### Representation - Present: Focus on technical capabilities - Improve: Diverse use cases and developer scenarios #### Psychological safety - Present: Non-judgmental approach to AI assistance - Improve: Address concerns about AI replacing human creativity ### ✅ Logical Validity #### Strong reasoning - Deductive strength: Logical progression from centralized management to improved workflow - Inductive strength: Examples support claims about workflow improvements - Causal reasoning: Clear connection between parallel processing and efficiency gains - Analogical strength: Effective comparison to traditional sequential workflows #### Fallacies - Appeal to novelty: Assumes newer multi-agent approach is inherently better - Hasty generalization: Limited evidence for broad productivity claims #### Tone of voice - Overall tone: Professional and enthusiastic, focused on practical benefits ### 🛡️ DARVO Analysis - Deny: No denial patterns present - Attack: No attacks on existing approaches - Reverse victim/offender: Not applicable - Emotional cues and rhetoric: Positive framing emphasizing empowerment and efficiency ## 🎯 Conclusion GitHub's Agent HQ represents a significant evolution in AI-assisted development, moving from single-agent workflows to orchestrated multi-agent collaboration. The integration across web, mobile, and IDE environments suggests this will become a standard part of developer workflows. Success will depend on developers learning effective agent selection strategies and managing the complexity of parallel AI assistance. ## 🧱 Refactor Appendix ### 📚 Literature Split Pad Purpose: Stage CEE units that become Literature notes. Split level: H3 ### [H1] Multi-agent AI orchestration transforms development workflows > [!quote] [Use Claude, Codex and Copilot together on GitHub](https://www.youtube.com/watch?v=GuTQDXKwdJQ) by [[GitHub]] via [[YouTube]]. Published 2024. Accessed 2024-12-19. > "shift your development tasks from sequential to simultaneous with the agent of your choice" Demonstrates fundamental shift in AI-assisted coding from single to parallel agent usage. [view highlight](https://www.youtube.com/watch?v=GuTQDXKwdJQ#H1) #ai-tools #workflow #articleseed ID:rw_gh001 ### [H2] Centralized agent management enables unified AI workflow control > [!quote] [Use Claude, Codex and Copilot together on GitHub](https://www.youtube.com/watch?v=GuTQDXKwdJQ) > "manage all your cloud agent session from one centralized location on GitHub" Single interface for orchestrating multiple AI models reduces context switching overhead. [view highlight](https://www.youtube.com/watch?v=GuTQDXKwdJQ#H2) #developer-experience #productivity ID:rw_gh002 --- ### 🔖 Atomic Split Pad (Zettel Candidates) Purpose: Promote reusable claims into Atomic notes. Split level: H4 #### [Z1] Parallel AI agent assignment improves development velocity - Applies to projects using multiple AI coding assistants simultaneously - Pointer to 00:00 timestamp demonstration - Links: [[AI-Assisted-Development]] [[Workflow-Optimization]] #### [Z2] Mobile agent management extends AI workflow accessibility - Enables remote task assignment and monitoring capabilities - Pointer to 00:43 mobile demonstration - Links: [[Remote-Development]] [[Mobile-Productivity --- ## Full Transcript [00:00] With GitHub Copilot, you can shift your<br>[00:02] development tasks from sequential to<br>[00:04] simultaneous with the agent of your<br>[00:06] choice in agent HQ. The agents tab on<br>[00:09] GitHub is where you can manage all your<br>[00:11] cloud agent session. From here, you can<br>[00:12] assign ad hoc tasks to copilot to claw<br>[00:16] byanthropic to codeex or to a custom<br>[00:19] agent. We can also use this tool to<br>[00:21] monitor our tasks all from one<br>[00:23] centralized location on GitHub. Now most<br>[00:26] change requests come via issues and you<br>[00:28] can assign issues directly to an agent<br>[00:31] like copilot codeex or claude. Choose<br>[00:34] the branch you wish to use as the base<br>[00:36] and again the agent you want and let the<br>[00:38] agent do the rest. Inspiration can<br>[00:41] strike from anywhere including away from<br>[00:43] your desktop. We can manage agents and<br>[00:46] even assign tasks right from GitHub<br>[00:48] mobile. Given that developers spend the<br>[00:51] bulk of their time in an IDE, it's<br>[00:54] natural to expect to have access to the<br>[00:56] same functionality from there. In VS<br>[00:59] Code, you can orchestrate all your agent<br>[01:01] sessions local and cloud from your IDE.<br>[01:04] Just like before, we can assign to<br>[01:06] claude or codeex or to copilot using a<br>[01:09] custom agent or specified model. These<br>[01:12] tools allow you to assign and manage<br>[01:14] parallel coding agent sessions all<br>[01:16] within your existing workflow. <!-- deep-read:sections end --> ## Highlights - **With GitHub Copilot, you can** **shift your development tasks** **from sequential to simultaneous** **with the agent of** **your choice in Agent HQ.** **The agents tab on GitHub** **is where you can manage** **all your cloud agent session.** **From here, you can assign** **ad hoc tasks to Copilot,** **to Claude by Anthropic,** **to Codex,** **or to a custom agent.** **We can also use this** **tool to monitor our tasks,** **all from one** **centralized location on GitHub.** **Now, most change requests come via issues** **and you can assign** **issues directly to an agent** ([View Highlight](https://read.readwise.io/read/01khfjbx81c6f8c30s4wmj6mw9)) - Tags: 👻 ai highlighted ## New highlights added February 14, 2026 at 11:24 PM - **With GitHub Copilot, you can** **shift your development tasks** **from sequential to simultaneous** **with the agent of** **your choice in Agent HQ.** **The agents tab on GitHub** **is where you can manage** **all your cloud agent session.** **From here, you can assign** **ad hoc tasks to Copilot,** **to Claude by Anthropic,** **to Codex,** **or to a custom agent.** **We can also use this** **tool to monitor our tasks,** **all from one** **centralized location on GitHub.** **Now, most change requests come via issues** **and you can assign** **issues directly to an agent** ([View Highlight](https://read.readwise.io/read/01khfjbx81c6f8c30s4wmj6mw9)) - Tags: 👻 ai highlighted ## New highlights added February 15, 2026 at 12:24 AM - **With GitHub Copilot, you can** **shift your development tasks** **from sequential to simultaneous** **with the agent of** **your choice in Agent HQ.** **The agents tab on GitHub** **is where you can manage** **all your cloud agent session.** **From here, you can assign** **ad hoc tasks to Copilot,** **to Claude by Anthropic,** **to Codex,** **or to a custom agent.** **We can also use this** **tool to monitor our tasks,** **all from one** **centralized location on GitHub.** **Now, most change requests come via issues** **and you can assign** **issues directly to an agent** ([View Highlight](https://read.readwise.io/read/01khfjbx81c6f8c30s4wmj6mw9)) - Tags: 👻 ai highlighted ## New highlights added February 15, 2026 at 1:25 AM - **With GitHub Copilot, you can** **shift your development tasks** **from sequential to simultaneous** **with the agent of** **your choice in Agent HQ.** **The agents tab on GitHub** **is where you can manage** **all your cloud agent session.** **From here, you can assign** **ad hoc tasks to Copilot,** **to Claude by Anthropic,** **to Codex,** **or to a custom agent.** **We can also use this** **tool to monitor our tasks,** **all from one** **centralized location on GitHub.** **Now, most change requests come via issues** **and you can assign** **issues directly to an agent** ([View Highlight](https://read.readwise.io/read/01khfjbx81c6f8c30s4wmj6mw9)) - Tags: 👻 ai highlighted ## New highlights added February 15, 2026 at 2:24 AM - **With GitHub Copilot, you can** **shift your development tasks** **from sequential to simultaneous** **with the agent of** **your choice in Agent HQ.** **The agents tab on GitHub** **is where you can manage** **all your cloud agent session.** **From here, you can assign** **ad hoc tasks to Copilot,** **to Claude by Anthropic,** **to Codex,** **or to a custom agent.** **We can also use this** **tool to monitor our tasks,** **all from one** **centralized location on GitHub.** **Now, most change requests come via issues** **and you can assign** **issues directly to an agent** ([View Highlight](https://read.readwise.io/read/01khfjbx81c6f8c30s4wmj6mw9)) - Tags: 👻 ai highlighted ## New highlights added February 15, 2026 at 3:24 AM - **With GitHub Copilot, you can** **shift your development tasks** **from sequential to simultaneous** **with the agent of** **your choice in Agent HQ.** **The agents tab on GitHub** **is where you can manage** **all your cloud agent session.** **From here, you can assign** **ad hoc tasks to Copilot,** **to Claude by Anthropic,** **to Codex,** **or to a custom agent.** **We can also use this** **tool to monitor our tasks,** **all from one** **centralized location on GitHub.** **Now, most change requests come via issues** **and you can assign** **issues directly to an agent** ([View Highlight](https://read.readwise.io/read/01khfjbx81c6f8c30s4wmj6mw9)) - Tags: 👻 ai highlighted ## New highlights added February 15, 2026 at 4:24 AM - **With GitHub Copilot, you can** **shift your development tasks** **from sequential to simultaneous** **with the agent of** **your choice in Agent HQ.** **The agents tab on GitHub** **is where you can manage** **all your cloud agent session.** **From here, you can assign** **ad hoc tasks to Copilot,** **to Claude by Anthropic,** **to Codex,** **or to a custom agent.** **We can also use this** **tool to monitor our tasks,** **all from one** **centralized location on GitHub.** **Now, most change requests come via issues** **and you can assign** **issues directly to an agent** ([View Highlight](https://read.readwise.io/read/01khfjbx81c6f8c30s4wmj6mw9)) - Tags: 👻 ai highlighted ## New highlights added February 15, 2026 at 5:24 AM - **With GitHub Copilot, you can** **shift your development tasks** **from sequential to simultaneous** **with the agent of** **your choice in Agent HQ.** **The agents tab on GitHub** **is where you can manage** **all your cloud agent session.** **From here, you can assign** **ad hoc tasks to Copilot,** **to Claude by Anthropic,** **to Codex,** **or to a custom agent.** **We can also use this** **tool to monitor our tasks,** **all from one** **centralized location on GitHub.** **Now, most change requests come via issues** **and you can assign** **issues directly to an agent** ([View Highlight](https://read.readwise.io/read/01khfjbx81c6f8c30s4wmj6mw9)) - Tags: 👻 ai highlighted ## New highlights added February 15, 2026 at 6:24 AM - **With GitHub Copilot, you can** **shift your development tasks** **from sequential to simultaneous** **with the agent of** **your choice in Agent HQ.** **The agents tab on GitHub** **is where you can manage** **all your cloud agent session.** **From here, you can assign** **ad hoc tasks to Copilot,** **to Claude by Anthropic,** **to Codex,** **or to a custom agent.** **We can also use this** **tool to monitor our tasks,** **all from one** **centralized location on GitHub.** **Now, most change requests come via issues** **and you can assign** **issues directly to an agent** ([View Highlight](https://read.readwise.io/read/01khfjbx81c6f8c30s4wmj6mw9)) - Tags: 👻 ai highlighted ## New highlights added February 15, 2026 at 7:24 AM - **With GitHub Copilot, you can** **shift your development tasks** **from sequential to simultaneous** **with the agent of** **your choice in Agent HQ.** **The agents tab on GitHub** **is where you can manage** **all your cloud agent session.** **From here, you can assign** **ad hoc tasks to Copilot,** **to Claude by Anthropic,** **to Codex,** **or to a custom agent.** **We can also use this** **tool to monitor our tasks,** **all from one** **centralized location on GitHub.** **Now, most change requests come via issues** **and you can assign** **issues directly to an agent** ([View Highlight](https://read.readwise.io/read/01khfjbx81c6f8c30s4wmj6mw9)) - Tags: 👻 ai highlighted ## New highlights added February 15, 2026 at 8:24 AM - **With GitHub Copilot, you can** **shift your development tasks** **from sequential to simultaneous** **with the agent of** **your choice in Agent HQ.** **The agents tab on GitHub** **is where you can manage** **all your cloud agent session.** **From here, you can assign** **ad hoc tasks to Copilot,** **to Claude by Anthropic,** **to Codex,** **or to a custom agent.** **We can also use this** **tool to monitor our tasks,** **all from one** **centralized location on GitHub.** **Now, most change requests come via issues** **and you can assign** **issues directly to an agent** ([View Highlight](https://read.readwise.io/read/01khfjbx81c6f8c30s4wmj6mw9)) - Tags: 👻 ai highlighted ## New highlights added February 15, 2026 at 12:03 PM - **With GitHub Copilot, you can** **shift your development tasks** **from sequential to simultaneous** **with the agent of** **your choice in Agent HQ.** **The agents tab on GitHub** **is where you can manage** **all your cloud agent session.** **From here, you can assign** **ad hoc tasks to Copilot,** **to Claude by Anthropic,** **to Codex,** **or to a custom agent.** **We can also use this** **tool to monitor our tasks,** **all from one** **centralized location on GitHub.** **Now, most change requests come via issues** **and you can assign** **issues directly to an agent** ([View Highlight](https://read.readwise.io/read/01khfjbx81c6f8c30s4wmj6mw9)) - Tags: 👻 ai highlighted