# Persistent Coding Agent Workflows and the Infra Forming Around Them *By Coding Agents Alpha Tracker • May 2, 2026* The practical shift today is from one-shot coding help to persistent agent systems: daily loops, agent-native install prompts, portable context in Codex, and new infra for heavy agent workflows. This brief pulls out the concrete patterns and releases you can actually use. ## 🔥 TOP SIGNAL Today’s real edge is **persistent agent workflow design**, not another model leaderboard [^1][^2]. Alexander Embiricos says OpenAI growth teammate Sahil Punamia’s internal "Lord Bottleneck" started as separate Codex-assisted steps and became a daily loop that reviews past experiments, proposes new ones, generates code/config after the team picks, and runs again [^3][^1]. Karpathy makes the matching interface point from the tooling side: specify the *outcome*, let the agent adapt to the local machine, and let it debug setup in the loop [^2]. ## ⚡ TRY THIS - **Turn repeated work into a morning agent loop.** Embiricos’ "Lord Bottleneck" pattern is straightforward: start by using Codex on each subtask separately—data analysis, experiment ideation, code generation, running the experiment, results analysis, deck writing—then stitch those steps into one reusable skill, then tell it to run every morning [^1]. The durable pattern is the important part: don’t start with full automation; chain together the steps that already work [^3][^1]. - **Write install docs as a prompt, not a shell script.** Karpathy’s OpenClaw example: instead of shipping a giant cross-platform installer, publish a copy-paste prompt that tells the agent the desired outcome and available tools; the agent can inspect the environment, handle platform differences, and debug setup itself [^2]. For Here Now, the whole install flow was effectively: > "I'd like you to set up here now the web hosting and cloud storage service for agents install as a skill if I have npm and if not, do this instead." [^2] If you build dev tools, Karpathy’s broader complaint is worth taking literally: docs should answer "what is the thing I should copy paste to my agent?" [^2] - **Prototype in two stages: data pipeline first, UI prompt second.** Simon Willison built an iNaturalist viewer entirely on his phone with Claude Code for web: first he created a small Python CLI to fetch and "clump" observations; then he ran that in a git-scraping repo to emit `clumps.json`; only then did he prompt for the frontend [^4]. His exact UI prompt was: > `Build inat-sightings.html - an app that does a fetch() against https://raw.githubusercontent.com/simonw/inaturalist-clumps/refs/heads/main/clumps.json and then displays all of the observations on one page using the https://static.inaturalist.org/photos/538073008/small.jpg small.jpg URLs for the thumbnails - with loading=lazy - but when a thumbnail is clicked showing the large.jpg in an HTML modal. Both small and large should include the common species names if available` [^4] - **If you script Claude Code, watch your recent commit messages.** Theo highlighted Claude Code’s programmatic `-p` prompt mode and appended system prompts for automation, but warned that recent commit history mentioning tools like OpenClaw or `Hermes MD` could trigger refusals or extra billing, even in an empty repo, based on his demo [^5][^6][^5]. His practical warning was simple: be careful what you put in commit messages when using Claude Code [^7]. ## 📡 WHAT SHIPPED - **Crabbox 0.1.0** — Peter Steinberger’s answer to "too many agents, too many test suites": remote Linux test boxes on AWS/Hetzner, dirty checkout sync, warm boxes with friendly slugs, and idle auto-free. Install with `brew install openclaw/tap/crabbox`. Site: [crabbox.sh](https://crabbox.sh) [^8] - **Codex import flow** — OpenAI added migration from other agents: import projects, settings, plugins, agents, and project configuration "in just a few clicks," with Romain Huet framing it as a seamless move to Codex [^9][^10][^9][^10] - **Codex pets** — New `/pet` feature for persistent context. Tibo says it makes him more productive because the context follows him while multitasking, and Riley Brown says his agent could still write to a pet notebook created four days earlier [^11][^12][^13] - **Devin inside the shell** — Cognition added shell integration: run `devin shell setup`, hit `Ctrl+G`, and Devin can see the current terminal screen to help in place [^14] - **`llm-openai-via-codex 0.1a0`** — Simon Willison released a plugin that reuses Codex CLI credentials for API calls via `llm`. Release: [llm-openai-via-codex 0.1a0](https://github.com/simonw/llm-openai-via-codex/releases/tag/0.1a0) [^15] - **GPT-5.5 migration guidance for coding agents** — OpenAI says Codex and the main model line are unified, suggests treating GPT-5.5 as a fresh model family, and provides a Codex-side migration path with `$openai-docs migrate this project to gpt-5.5`; for multi-step tasks, the agent should send a short user-visible update before tool calls [^15] ## 🎬 GO DEEPER - **21:35-23:27 — Karpathy on "vibe coding" vs "agentic engineering."** Best short framing in today’s source set: vibe coding raises the floor, agentic engineering raises the ceiling, and the job becomes coordinating fallible agents without dropping the quality bar [^2]. [![Reacting to "Why AI is so smart but also so dumb?"](https://img.youtube.com/vi/pngC-TH8M0U/hqdefault.jpg)](https://youtube.com/watch?v=pngC-TH8M0U&t=1294) *Reacting to "Why AI is so smart but also so dumb?" (21:34)* - **20:51-24:14 — Theo on why commit history can leak into Claude Code behavior.** Worth watching if you wrap Claude Code in harnesses or scripts: his demo argues recent git history gets surfaced in a way that can affect behavior and billing [^5]. [![Be careful what you name your markdown files...](https://img.youtube.com/vi/J8O9LLpJNrg/hqdefault.jpg)](https://youtube.com/watch?v=J8O9LLpJNrg&t=1250) *Be careful what you name your markdown files... (20:50)* - **Study the tiny-but-complete prototype chain.** Simon’s [simonw/inaturalist-clumper](https://github.com/simonw/inaturalist-clumper) plus [simonw/inaturalist-clumps](https://github.com/simonw/inaturalist-clumps) is a clean pattern: build a small data-collection CLI, turn it into continually refreshed JSON, then prompt the agent for the UI on top [^4] - **Read the release surfaces, not just the screenshots.** If you want inspectable details, start with [Codex CLI 0.128.0 release notes](https://github.com/openai/codex/releases/tag/rust-v0.128.0), [LLM 0.32a0 changelog](https://llm.datasette.io/en/latest/changelog.html#a0-2026-04-28), and [crabbox.sh](https://crabbox.sh) [^15][^8] *Editorial take: the teams pulling ahead are not just picking better models; they’re turning agents into repeatable loops with durable context and infrastructure that can survive real work.* [^1][^12][^8][^2] --- ### Sources [^1]: [𝕏 post by @tbpn](https://x.com/tbpn/status/2050336898435207307) [^2]: [Reacting to "Why AI is so smart but also so dumb?"](https://www.youtube.com/watch?v=pngC-TH8M0U) [^3]: [𝕏 post by @embirico](https://x.com/embirico/status/2050347447780811180) [^4]: [iNaturalist Sightings](https://simonwillison.net/2026/May/1/inat-sightings) [^5]: [Be careful what you name your markdown files...](https://www.youtube.com/watch?v=J8O9LLpJNrg) [^6]: [𝕏 post by @theo](https://x.com/theo/status/2049645973350363168) [^7]: [𝕏 post by @theo](https://x.com/theo/status/2050331062946197900) [^8]: [𝕏 post by @steipete](https://x.com/steipete/status/2050139774451671285) [^9]: [𝕏 post by @romainhuet](https://x.com/romainhuet/status/2050352052128317549) [^10]: [𝕏 post by @OpenAI](https://x.com/OpenAI/status/2050290618187055175) [^11]: [𝕏 post by @OpenAIDevs](https://x.com/OpenAIDevs/status/2050275713824211041) [^12]: [𝕏 post by @thsottiaux](https://x.com/thsottiaux/status/2050280111820406795) [^13]: [𝕏 post by @rileybrown](https://x.com/rileybrown/status/2050370477475106958) [^14]: [𝕏 post by @cognition](https://x.com/cognition/status/2050268727997022498) [^15]: [DeepSeek v4, and the end of the OpenAI/Microsoft AGI clause](https://simonw.substack.com/p/deepseek-v4-and-the-end-of-the-openaimicrosoft)