The final piece is a merge queue: automatically merge agent-generated PRs once they’re up to date with main, fully tested, and approved.
https://t.co/GANQjlSKda
Linear is on its way to becoming a fully automated software factory. The next logical step is to automatically review every PR, feed the requested changes back to the original agent, and let the agents iterate until the task converges. I made a review-quill for that. https://t.co/8NSNuXTRev
Introducing coding sessions.
Linear Agent can now triage issues, investigate the cause, write the fix, open a PR, and bring the code back for review.
All shared with your team in Linear.
It’s already possible to connect Codex to Linear as an agent. It requires some setup with API keys and webhooks, but it lets you use your existing Codex subscription. Check out PatchRelay as an example. https://t.co/FFO9f7Gflz
I can imagine a lot of workflows around the codex app server: queue up tasks overnight in Linear, hand them off to an agent like PatchRelay, and let it pair with ReviewQuill to push a feature to prod through a chain of batched coding & review tasks. By morning, everything’s ready and the Codex limits barely moved. Beautiful.
6/ Give Claude a way to verify its work
Finally, make sure Claude has a way to verify its work. This has always been a way to 2-3x what you get out of Claude, and with 4.7 it's more important than ever.
Verification looks different depending on the task. For backend work, make sure Claude knows how to start up your server/service to test it end to end; for frontend work, use the Claude Chromium extension to give Claude a way to control your browser; for desktop apps, use computer use.
Personally, many of my prompts these days look like "Claude do blah blah /go". /go is a skill that has Claude
1. Test itself end to end using bash, browser, or computer use
2. Run the /simplify skill
3. Put up a PR
For long running work, verification is important because that way when you come back to a task, you know the code works.
@erratlinezero@mitchellh You already can use https://t.co/oGItIkoXzm
Based on libghostty, works awesome together with auto-ssh: long lived sessions with reconnects.
Introducing cmux: the open-source terminal built for coding agents.
- Vertical tabs
- Blue rings around panes that need attention
- Built-in browser
- Based on Ghostty
When Claude Code needs you, the pane glows blue and the sidebar tells you why.
No Electron/Tauri. Just Swift/Appkit.
@OlgaBeshley Третье занятие - слишком рано чтобы предъявлять какие либо требования.
Мозг ещё только сопоставляет что там было в учебнике и пытается понять как автомобиль реагирует.
Уверенность движений и быстрота реакции придёт с практикой. Проблема не в тебе, её вообще нет.
@OlgaBeshley Мозгу нужно дать время освоится за рулём – он сам поймёт на что нужно обращать внимание, с какой силой давить на педали и в какой момент крутить руль.
Инструктора нужно подобрать такого который будет создавать условия подходящие по уровню - немного сложно, но не стресс.
@0xfe0d Мне LanguageTransfer очень зашел для изучения испанского. Курс хорош своей методологией: он не просто даёт контент, он выстраивает способ мышления для освоения языка. Этого нет в сове.
С греческим не сталкивался, но курс от того же автора-полиглота. Try before you deny.
@irvinebroque@CloudflareDev I actually wanted my local dev to fully mirror prod. I have one worker with a queue and a workflow: when an event hits the queue, the handler runs and should trigger the workflow. In prod it works, locally queue works, but the workflow never starts: no logs, no DB changes.