{
  "id": 13572579,
  "title": "One inbox for all my agents",
  "url": "https://urgent.news/2026/10/10/one-inbox-for-all-my-agents",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-10T23:11:28.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/arielf/one-inbox-for-all-my-agents-42lk"
  },
  "original_language": "en",
  "account": "Managing multiple coding agents simultaneously can be a challenge, with each agent running in its own terminal and requiring manual attention to keep track of their progress. The author, however, found a solution in handup, a tool that unifies all these agents into a single inbox.\n\nThis inbox, accessible on both desktop and mobile devices, consolidates all the requests, decisions, and completed tasks from each agent. It saves the author 1,700 instances of searching through terminals, as each request, update, question, review, and go/no-go call now appears in one place. The author can respond to these items instantly, whether it's approving a deploy, answering a question, or reviewing a draft.\n\nResponses are direct and result in the agent continuing its work. If follow-up is needed, the author can simply type a response under the original request, and the agent picks it up as the next task. The inbox also serves as a log of the agents' decisions, with the author's feedback serving as the rationale behind each action.\n\nOver time, the inbox not only becomes a communication hub but also a knowledge base. When the author denies a request, they provide a reason, which the agent receives along with the denial. If the same kind of decision is made twice, the answer is stored in the agent's skill set or the project's AGENTS.md, enabling the agent to handle similar decisions on its own in the future.\n\nThe author's ultimate goal with handup is to reduce the number of decisions needed, as each decision recorded in the inbox helps the agents become more autonomous. The inbox acts as a feedback loop, where each decision either unblocks an agent or teaches it not to ask the same question again.\n\nhandup utilizes MCP tools such as notify for updates, ask_question for questions requiring human input, and request_approval for consequential actions. The agents receive instructions on when to use these tools through handup prompt. The entire setup is as simple as starting handup, connecting every detected agent, and printing the instruction for AGENTS.md or CLAUDE.md.\n\nBy simplifying the process of managing multiple agents, handup frees up time for tasks that truly require human intervention, making the entire coding process more efficient and less time-consuming.",
  "summary": "I usually have a handful of coding agents going at once: Claude Code in one terminal, Codex in another, a couple more on a server. They all run with permission prompts off ( --dangerously-skip-permissions for Claude Code, --yolo for Codex), so they move fast. The slow part was me, flipping through terminals to find which agent was waiting, which had finished, and which needed a decision. Now the…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}