--- title: "AgentRQ Blog | Claude Code & AI Agent Insights" url: https://agentrq.com/blog/ description: "Technical insights on Claude Code, MCP, and building AI agents with human-in-the-loop workflows." --- AgentRQ Blog # Technical Insights Claude Code, MCP, and building AI agents with human-in-the-loop workflows. October 2, 2026 | AgentRQ Team ## [What Is YOLO Mode? How to Enable It in Every Agent](https://agentrq.com/blog/what-is-yolo-mode-and-how-to-turn-it-on-in-every-coding-agent) YOLO mode lets a coding agent act without asking. The exact flag in Claude Code, Codex, Antigravity, Gemini CLI, Copilot, Cline and OpenCode, and safe use. [Read Post](https://agentrq.com/blog/what-is-yolo-mode-and-how-to-turn-it-on-in-every-coding-agent) October 2, 2026 | AgentRQ Team ## [How ACP Gateway Finds Agents and Their Models](https://agentrq.com/blog/how-acp-gateway-discovers-acp-compatible-agents-and-their-models) How AgentRQ's ACP Gateway finds every ACP-compatible agent your machine can run, and every model each one offers, with no list kept by hand. [Read Post](https://agentrq.com/blog/how-acp-gateway-discovers-acp-compatible-agents-and-their-models) October 2, 2026 | AgentRQ Team ## [Autonomous Agents vs. Agentic Workflows, Explained](https://agentrq.com/blog/autonomous-agents-vs-agentic-workflows-the-real-difference) Autonomous agents and agentic workflows answer two questions: who decides the next step, and who approves it. Anthropic's framing, a table and how to pick. [Read Post](https://agentrq.com/blog/autonomous-agents-vs-agentic-workflows-the-real-difference) October 2, 2026 | AgentRQ Team ## [Which Coding Agents Support ACP?](https://agentrq.com/blog/which-coding-agents-support-acp-agent-client-protocol) Which coding agents support ACP? A support table for 33 agents, including Claude Code, Codex, Gemini CLI, Cursor and OpenCode, with each one's ACP command. [Read Post](https://agentrq.com/blog/which-coding-agents-support-acp-agent-client-protocol) September 28, 2026 | AgentRQ Team ## [AgentRQ for Chrome: Your Agents Can Now Use Websites](https://agentrq.com/blog/agentrq-chrome-extension-webmcp-bridge-for-local-and-remote-agents) The AgentRQ Chrome extension is live: Claude Code terminals in your toolbar, and a bridge that hands any site's WebMCP tools to local or remote agents. [Read Post](https://agentrq.com/blog/agentrq-chrome-extension-webmcp-bridge-for-local-and-remote-agents) September 28, 2026 | AgentRQ Team ## [One Workspace, One Job: Why I Isolate My AI Agents](https://agentrq.com/blog/why-i-use-isolated-single-purpose-workspaces-for-ai-agents) Why I run every AI agent in an isolated workspace that does one job: memory that stays true, a smaller starting context, and the fair counterarguments. [Read Post](https://agentrq.com/blog/why-i-use-isolated-single-purpose-workspaces-for-ai-agents) September 27, 2026 | AgentRQ Team ## [Every Task Now Keeps Its Own Timeline](https://agentrq.com/blog/task-status-history-timeline-for-ai-agent-tasks) AgentRQ now records every task status change and draws it as a timeline: time worked, time blocked, time waiting on you, and start to close. [Read Post](https://agentrq.com/blog/task-status-history-timeline-for-ai-agent-tasks) September 27, 2026 | AgentRQ Team ## [Task Latency: How Long Your Agents' Tasks Take](https://agentrq.com/blog/task-latency-statistics-for-ai-agent-workspaces) AgentRQ stats pages now chart how long closed tasks take, per workspace and across your account: start to close, worked, blocked and waiting on you. [Read Post](https://agentrq.com/blog/task-latency-statistics-for-ai-agent-workspaces) September 24, 2026 | AgentRQ Team ## [You Fixed the Playbook on Your Laptop. Nine Other Agents Never Heard.](https://agentrq.com/blog/distributed-agent-skills-over-mcp-across-workspaces) Agent skills used to live in a directory on one machine. AgentRQ v0.8.0 serves them over MCP, so every agent loads the same current playbook — and can improve it. [Read Post](https://agentrq.com/blog/distributed-agent-skills-over-mcp-across-workspaces) September 20, 2026 | AgentRQ Team ## [Jev and System One Models: A New Kind of AI Model](https://agentrq.com/blog/jev-system-one-models) TypeSafe AI's Jev skips text generation, outputting typed, calibrated decisions in milliseconds. What it is, and why fast structured decisions matter for agents. [Read Post](https://agentrq.com/blog/jev-system-one-models) September 19, 2026 | AgentRQ Team ## [Start Every Task on a Clean Context, and Stop Paying for the Last One](https://agentrq.com/blog/clear-context-before-every-task) AgentRQ sends /clear to a Claude Code agent before it picks a task up, so every task starts on a clean context instead of re-reading the last one on every turn. [Read Post](https://agentrq.com/blog/clear-context-before-every-task) September 17, 2026 | AgentRQ Team ## [Machines: Run Agents on Hardware You Own, and Watch the Terminal From Anywhere](https://agentrq.com/blog/run-agents-on-your-own-machines) AgentRQ v0.7.0 adds Machines — install agentrqd on a computer you own, enrol it once, and start a coding agent on it from the control panel, then watch its terminal and type into it from anywhere. [Read Post](https://agentrq.com/blog/run-agents-on-your-own-machines) September 17, 2026 | AgentRQ Team ## [MCP Attachment Base64 Cost: Why AgentRQ Has a Workspace CLI](https://agentrq.com/blog/base64-cost-of-mcp-attachments) What an MCP attachment costs in base64: the encoded file sits in the agent's context for the whole session. The AgentRQ workspace CLI keeps it out. [Read Post](https://agentrq.com/blog/base64-cost-of-mcp-attachments) September 6, 2026 | AgentRQ Team ## [Your Agents Keep Forgetting. Now the Workspace Remembers.](https://agentrq.com/blog/workspace-memory) Everything your agent learns dies when the task ends. Workspace memory gives every agent you run — Claude, Gemini, Codex, on any machine — one shared set of notes that outlives the task, and lets you read them. [Read Post](https://agentrq.com/blog/workspace-memory) September 5, 2026 | AgentRQ Team ## [WebMCP: Handing the Interface to an Agent in Your Browser](https://agentrq.com/blog/webmcp-browser-agents) AgentRQ now hands its whole interface to an AI agent running in your browser as 52 typed WebMCP tools — no install, no configuration, no token. Here is exactly what shipped and how to check it. [Read Post](https://agentrq.com/blog/webmcp-browser-agents) September 5, 2026 | AgentRQ Team ## [How to Connect and Use Antigravity with AgentRQ](https://agentrq.com/blog/connect-antigravity-agentrq) A step-by-step guide to connecting Google Antigravity to an AgentRQ workspace with the open-source ACP Gateway — sign-in, workspace setup, .mcp.json, Antigravity login, choosing a Gemini model, and running the bridge. [Read Post](https://agentrq.com/blog/connect-antigravity-agentrq) August 29, 2026 | AgentRQ Team ## [A Few Seconds to Reconsider Before a Message Sends](https://agentrq.com/blog/message-send-delay) A new per-workspace setting holds outgoing chat messages for a few seconds before delivery, with Send Now and Cancel controls — a brief buffer to catch a typo or reconsider before an agent acts on it. [Read Post](https://agentrq.com/blog/message-send-delay) August 24, 2026 | AgentRQ Team ## [Moving Tasks Between Workspaces](https://agentrq.com/blog/move-task-between-workspaces) Tasks created in the wrong workspace used to mean starting over. Right-click a task, pick a destination, and it moves with its full chat history, tool calls, and attachments intact. [Read Post](https://agentrq.com/blog/move-task-between-workspaces) August 23, 2026 | AgentRQ Team ## [Testing and Applying MCP Server Fixes With an Independent Conformance Suite](https://agentrq.com/blog/mcp-server-conformance-testing) We ran an independent MCP conformance suite against our Core and Workspace MCP servers, fixed the real bugs it found, and wrote tests pinning down the cases where it was simply wrong. Here's what that looked like in practice. [Read Post](https://agentrq.com/blog/mcp-server-conformance-testing) August 20, 2026 | AgentRQ Team ## [Team Collaboration on AgentRQ Tasks, Straight From Slack](https://agentrq.com/blog/slack-team-collaboration) AgentRQ's Slack integration turns a workspace into a shared room: one private channel, every teammate's own agent, and every approve/deny attributed to the person who clicked it. Here's how team collaboration works without anyone touching a shared login. [Read Post](https://agentrq.com/blog/slack-team-collaboration) August 18, 2026 | AgentRQ Team ## [Workflows Without Decision Trees](https://agentrq.com/blog/workflows-without-decision-trees) AgentRQ workflows have no if/else, no conditions, no decision nodes. Routing is a lookup on (workflow, event), branching is a choice of event name, and loops are rejected at save time. Here's the mechanism, tested against a live v0.4.1 build. [Read Post](https://agentrq.com/blog/workflows-without-decision-trees) August 17, 2026 | AgentRQ Team ## [Workflows: Turning Scattered Event Triggers Into a Pipeline You Can See](https://agentrq.com/blog/agentrq-workflows) AgentRQ v0.4.0 adds Workflows — chain events and workspaces into one named pipeline you can draw or type, review before it runs, and execute in isolation. Agents coordinate by publishing named events: no conditionals, no back-and-forth. [Read Post](https://agentrq.com/blog/agentrq-workflows) August 15, 2026 | AgentRQ Team ## [Tool Call History: A Trajectory View of Everything an Agent Did](https://agentrq.com/blog/tool-call-stack-view) Every tool call and message on a task now lives in one searchable trajectory view — a 3-lane minimap plus a master-detail panel with Summary, Payload, and Content tabs. [Read Post](https://agentrq.com/blog/tool-call-stack-view) August 15, 2026 | AgentRQ Team ## [Delegate to DeepSeek Harness From Anywhere: the New AgentRQ Plugin](https://agentrq.com/blog/deepseek-harness-plugin) A new AgentRQ plugin for DeepSeek Harness lets you assign tasks from the AgentRQ web app or your phone and have them pushed live into a running DeepSeek Harness agent — no polling, no leaving the harness. [Read Post](https://agentrq.com/blog/deepseek-harness-plugin) August 14, 2026 | AgentRQ Team ## [Local Speech-to-Text: Talk to Your Agent Instead of Typing](https://agentrq.com/blog/local-speech-to-text) Speak tasks instead of typing them: on a computer AgentRQ transcribes with a local Whisper model in your browser; on a phone, with its own dictation. [Read Post](https://agentrq.com/blog/local-speech-to-text) August 14, 2026 | AgentRQ Team ## [Local AI Task Titles: Your Browser Writes the Headline, Not a Server](https://agentrq.com/blog/local-ai-task-title-generation) AgentRQ can now write your task titles for you — using a small language model that runs entirely inside your browser, so the task description never leaves your machine. [Read Post](https://agentrq.com/blog/local-ai-task-title-generation) August 14, 2026 | AgentRQ Team ## [Workspace Activity Heatmaps: See When Your Agents Actually Work](https://agentrq.com/blog/workspace-activity-heatmaps) AgentRQ's workspace analytics page now includes GitHub-style contribution heatmaps for task completions and messages, adapting from hour-by-hour to day-by-day as you zoom out. [Read Post](https://agentrq.com/blog/workspace-activity-heatmaps) May 24, 2026 | AgentRQ Team ## [Running Gemma 4 31B with MTP on Nvidia NGX Spark via Claude Code](https://agentrq.com/blog/gemma4-31b-mtp) Step-by-step guide on running Gemma 4 31B with Multi-Token Prediction (MTP) on Nvidia NGX Spark using Docker and Claude Code. [Read Post](https://agentrq.com/blog/gemma4-31b-mtp) May 17, 2026 | AgentRQ Team ## [Expanding the Loop: Announcing Codex Support via the AgentRQ Codex Gateway](https://agentrq.com/blog/announcing-codex-gateway) Announcing the AgentRQ Codex Gateway. Connect OpenAI Codex and other ACP-compatible agents to your AgentRQ workspaces for remote human-in-the-loop task management. [Read Post](https://agentrq.com/blog/announcing-codex-gateway) May 17, 2026 | AgentRQ Team ## [Claude Channels Unlocked: Native Notifications for Everyone](https://agentrq.com/blog/claude-code-channels-unlocked) Anthropic has unlocked Claude Channels for all Claude Code users, removing the Claude Pro/Max subscription requirement. Here is what this means for API-based developers and the open-source community. [Read Post](https://agentrq.com/blog/claude-code-channels-unlocked) May 14, 2026 | AgentRQ Team ## [Agentic Task Management: The Bridge Between Humans and AI Agents](https://agentrq.com/blog/agentic-task-management-for-ai-agents) Discover how Agentic Task Management bridges the gap between autonomous AI agents and human oversight using AgentRQ. [Read Post](https://agentrq.com/blog/agentic-task-management-for-ai-agents) April 28, 2026 | AgentRQ Team ## [Closed-Loop vs. Open-Loop Agentic Systems: Why Feedback Is the Moat](https://agentrq.com/blog/closed-loop-agentic-systems) Closed-loop agentic systems continuously improve through feedback cycles. Open-loop systems execute and forget. Here's why the feedback loop is the single biggest determinant of which AI platforms win — and how AgentRQ is built around it. [Read Post](https://agentrq.com/blog/closed-loop-agentic-systems) April 28, 2026 | AgentRQ Team ## [Supervisor and Workers: The Architecture Behind Self-Growing Agentic Systems](https://agentrq.com/blog/supervisor-worker-agentic-system) AgentRQ's two-tier MCP architecture — a supervisor with global visibility and workers scoped to individual workspaces — is what makes closed-loop, self-growing multi-agent systems actually work in production. Here's how and why. [Read Post](https://agentrq.com/blog/supervisor-worker-agentic-system) April 24, 2026 | AgentRQ Team ## [DeepSeek-V4-Pro: 1.6 Trillion Parameters, MIT License, 1 Million Token Context](https://agentrq.com/blog/deepseek-v4-pro) DeepSeek-V4-Pro is a 1.6 trillion parameter open-source MoE model with MIT license, 1M token context, and benchmark results that challenge the best closed-source models. Here's what it actually is, how it works, and what it can do. [Read Post](https://agentrq.com/blog/deepseek-v4-pro) April 18, 2026 | AgentRQ Team ## [AgentRQ + ACP: Connecting Gemini CLI and Beyond to Your Remote Task Channel](https://agentrq.com/blog/agentrq-acp-gateway-gemini-cli) Learn how to connect Gemini CLI and other ACP-compatible agents to AgentRQ using the new open-source ACP Gateway. Remote task management for your local AI agents. [Read Post](https://agentrq.com/blog/agentrq-acp-gateway-gemini-cli) April 16, 2026 | AgentRQ Team ## [What Is an MCP Gateway — and Why Every AI Stack Needs One](https://agentrq.com/blog/mcp-gateway) MCP Gateways solve the hardest problems in connecting AI agents to real APIs: authentication, security, context overhead, and observability. Here's what they do and how HasMCP implements them. [Read Post](https://agentrq.com/blog/mcp-gateway) April 14, 2026 | AgentRQ Team ## [Agentic Workflows vs. Autonomous Workflows: What's the Difference and Why It Matters](https://agentrq.com/blog/agentic-workflows) Agentic workflows and autonomous workflows aren't the same thing. Here's how they differ, when each applies, and why the distinction matters for teams building with AI agents. [Read Post](https://agentrq.com/blog/agentic-workflows) April 10, 2026 | AgentRQ Team ## [Claude Mythos: The Frontier of AI-Driven Cybersecurity](https://agentrq.com/blog/claude-mythos-frontier-cybersecurity) Anthropic's unreleased Claude Mythos model has discovered thousands of zero-day vulnerabilities autonomously. Here is what we know about Mythos, its defensive role in Project Glasswing, and the future of frontier AI cybersecurity. [Read Post](https://agentrq.com/blog/claude-mythos-frontier-cybersecurity) April 6, 2026 | AgentRQ Team ## [Gemma 4 on iPhone: Frontier AI in Your Pocket, No Cloud Required](https://agentrq.com/blog/gemma-4-iphone-google-edge-gallery) Google's Gemma 4 now runs entirely on iPhone via the AI Edge Gallery app — with Thinking Mode and Agent Skills, fully offline. Here's why it matters for mobile AI, device control, and private intelligence. [Read Post](https://agentrq.com/blog/gemma-4-iphone-google-edge-gallery) April 4, 2026 | AgentRQ Team ## [Always-On Agent with Context vs. On-Demand Zero-Context Agent](https://agentrq.com/blog/always-on-vs-on-demand-agents) An always-on agent with workspace context outperforms an on-demand zero-context agent in almost every real project scenario. Here's why — and when each approach actually makes sense. [Read Post](https://agentrq.com/blog/always-on-vs-on-demand-agents) April 1, 2026 | AgentRQ Team ## [MolmoWeb: The Open-Source Web Agent That Navigates by Screenshots Alone](https://agentrq.com/blog/molmoweb-open-web-agent) Allen AI's MolmoWeb is a fully open 4B/8B web agent that navigates browsers using only screenshots — no DOM access, Apache 2.0 license. Here's how it works, what it actually beats, and how to run it. [Read Post](https://agentrq.com/blog/molmoweb-open-web-agent) April 1, 2026 | AgentRQ Team ## [The Portable AI Stack: M5 Max 128GB + Qwen3.5-35B + MolmoWeb](https://agentrq.com/blog/m5-max-128gb-local-ai-stack) M5 Max MacBook Pro with 128GB + Qwen3.5-35B + MolmoWeb = a self-contained AI stack that runs entirely offline. Here's what becomes possible when frontier-grade intelligence fits in a backpack. [Read Post](https://agentrq.com/blog/m5-max-128gb-local-ai-stack) April 1, 2026 | AgentRQ Team ## [TurboQuant: What Google's Extreme Compression Means for Local LLMs](https://agentrq.com/blog/turboquant-local-llms) Google's TurboQuant compresses LLM key-value caches to 3 bits with no accuracy loss — 8x faster on H100s. Here's how it works and what it means for running large models locally. [Read Post](https://agentrq.com/blog/turboquant-local-llms) March 30, 2026 | AgentRQ Team ## [MCP Tasks: How the Protocol Is Growing Up for Long-Running Agents](https://agentrq.com/blog/mcp-tasks-spec) MCP Tasks is an experimental spec that turns synchronous tool calls into durable state machines — poll for status, retrieve results later, and handle mid-task human input. Here's what it means for long-running AI agents. [Read Post](https://agentrq.com/blog/mcp-tasks-spec) March 28, 2026 | AgentRQ Team ## [ACP: The Protocol That Lets You Use Any Agent in Any Editor — and What Comes Next](https://agentrq.com/blog/acp-agent-client-protocol) ACP (Agent Client Protocol) is the open standard that turns any coding agent into a first-class IDE citizen. Here's what it enables today with stdio, and what becomes possible when remote transport arrives. [Read Post](https://agentrq.com/blog/acp-agent-client-protocol) March 28, 2026 | AgentRQ Team ## [What Claude Code Notification Channels Are Actually For](https://agentrq.com/blog/claude-notification-channels-use-cases) Claude Code notification channels flip the model from pull to push — here's what that unlocks, what the official channels cover, and how to think about what's worth wiring up. [Read Post](https://agentrq.com/blog/claude-notification-channels-use-cases) March 27, 2026 | AgentRQ ## [What Claude Computer Control Actually Is](https://agentrq.com/blog/claude-computer-control-explained) Claude Computer Control lets AI agents interact with screens, keyboards, and mice. Here's the technical breakdown of how it works, which models support it, and why it matters for agentic workflows. [Read Post](https://agentrq.com/blog/claude-computer-control-explained) March 26, 2026 | AgentRQ Team ## [How I Scale My Startup While Sleeping](https://agentrq.com/blog/scaling-my-startup-while-sleeping) How I use isolated AgentRQ workspaces for marketing, cold outreach, and coding — each scaling itself with human-in-the-loop workflow automations while I sleep. [Read Post](https://agentrq.com/blog/scaling-my-startup-while-sleeping) March 26, 2026 | AgentRQ Team ## [The Five Levels of AI Leverage](https://agentrq.com/blog/levels-of-ai-leverage) Most engineers have reached Level 3. Here's the full stack of AI leverage — from chatbots to fully autonomous, human-in-the-loop agentic workflows — and what it takes to reach the top. [Read Post](https://agentrq.com/blog/levels-of-ai-leverage) March 25, 2026 | AgentRQ Team ## [How Claude Code Channels Work](https://agentrq.com/blog/claude-code-channels) How Claude Code Channels push messages into a running session over MCP, how to build one, and every flag and setting, with a dated changelog. [Read Post](https://agentrq.com/blog/claude-code-channels)