What X is shipping with agents right now.

CyrilXBT@cyrilXBT·3h

A copy-paste settings.json plus two subagent files that split work by model: Opus plans, Sonnet codes, a Haiku scout searches files, and an Opus critic attacks the diff.

i can't comprehend why people run claude code on default settings one settings change and every model gets its own job: → opus plans every task and advises when stuck → sonnet 5.5 writes the actual code → haiku handles every file search → a critic agent tries to break your diff before you merge Opus thinks. The cheap models grind. paste this into claude code 👇 Add this to ~/.claude/settings.json. Keep every hook, permission and plugin I already have. Then create the two agent files below and show me the diff. ```json { "model": "opusplan", "advisorModel": "opus", "effortLevel": "medium", "modelSettings": { "claude-opus-5-5": { "effortLevel": "high" }, "claude-sonnet-5-5": { "effortLevel": "medium" } }, "env": { "CLAUDE_CODE_SUBAGENT_MODEL": "haiku" } } ``` ~/.claude/agents/scout.md ``` --- name: scout description: finds code fast. returns file paths and line numbers only. model: haiku tools: Read, Grep, Glob --- Locate exactly what was asked. Return file:line. No fixes. No opinions. ``` ~/.claude/agents/critic.md ``` --- name: critic description: reviews a diff or plan before any task is called done model: opus tools: Read, Grep, Bash --- Assume this is broken. Find how it fails. One line per issue, worst first. ```

CyrilXBT@cyrilXBT·6h

Eight repos that level up Claude Code: gstack, claude-flow, oh-my-claudecode, orca, claude-mem, context-mode, agent-skills, everything-claude-code.

Claude Code can build full apps. These 8 repos make it build them better. PLAN IT gstack → Garry Tan's setup that gives Claude Code roles like founder, eng manager and QA https://t.co/G2C5IJQq6O STAFF IT claude-flow → runs swarms of agents working together https://t.co/eQQsm7YfI4 oh-my-claudecode → multi agent orchestration on top of Claude Code https://t.co/MM6wevLBRt orca → parallel agents https://t.co/jysYX2z7T7 REMEMBER IT claude-mem → memory that survives between sessions https://t.co/1XI7Odux4C context-mode → keeps tool output from flooding the context https://t.co/QZh1fkM8RD DO IT RIGHT agent-skills → senior engineer habits: specs, TDD, reviews https://t.co/bONo0s9NJu everything-claude-code → agents, skills, hooks and configs in one place https://t.co/2Er9PREAih Bookmark this.

OpenAI Developers@OpenAIDevs·12h

OpenAI's playbook for pointing Codex at maintenance: it finds dead code and stale experiments, checks them against production telemetry, and drafts cleanup PRs.

Here’s how our engineers use Codex at OpenAI to dig through our codebase, find what needs attention, and chip away at the maintenance backlog: https://developers.openai.com/blog/maintaining-software-with-the-gpt-6-family

Pratham@Prathkum·14h

How the Claude Code team reviews agent PRs: one agent group hunts bugs, another votes on whether they're real, and a bot approval never replaces a human owner.

Anthropic published how the Claude Code team does code review. Claude writes about 90% of Claude Code. In their monorepo, changes per day went from 2,500 in May to over 30,000 in September. So agents write the code, and then agents review the PRs. One group hunts for bugs, another group votes on whether each bug is real. One PR went through 50+ rounds of review and fixes overnight, an agent author and an agent reviewer just feeding each other until a human pulled the plug. 😂 Still, merged PRs more than doubled since June and reverts have dropped by over 40%. And a bot’s approval never counts as a human’s. Whoever prompted the agent owns the change. I wonder why we even need separate agents to review code when agents are going to write the code anyway. Crazy!

ClaudeDevs@ClaudeDevs·15h

The Claude Code team's own write-up on reviewing agent-written PRs at volume, with numbers (2x merged PRs, 40% fewer reverts) and steps to adopt it.

Claude writes about 90% of Claude Code's code. From June to September, merged PRs more than doubled while the share reverted fell by more than 40%. We wrote about how we changed code review to get there, and how to adopt it: https://claude.dev/blog/solving-code-review-at-anthropic/

lauren@poteto·17h

The trust pyramid: put most effort into codebase design and static checks, use skills and AGENTS.md prose sparingly, and save human review for the highest-leverage spots.

how do you trust your agents' work more? copy this post/link and give it to your @bot or agent! 1. codebase improvements: this is just good software engineering - better architecture, stronger types and invariants, clear boundaries. agents love to copy what exists and take the shortest path, so make the easy thing the right thing. categorically eliminate problems through code and you won't need to think about it again. 2. static analysis: lint rules, tests, CI, verification (eg TLA+). very token efficient ways to catch issues ahead of time. 3. prose: comments, rules, AGENTS.md, skills - anything that can't be enforced and requires agent judgment. use it minimally, but do use /create-verification-skill. 4. human review: the hardest to scale. at the rate agents land code, it's pretty much impossible to thoroughly review every PR, so spend that effort on the highest leverage point, designing an environment where agents do the right thing by default. this forms an inverse pyramid: you want most of your effort at the top, least at the bottom. codebase improvements are deterministic, making it more token efficient, and lets you sleep more soundly at night. install https://t.co/u9xIlN3eN0 and create a new Project: 1. look through my slack/chat messages and code review comments where i've given feedback on PRs, code, architecture 2. use /recall on my past transcripts to learn how i steer agents 3. /poteto-mode use /correct and /architect on past commits to find the most common pitfalls agents fall into, then propose a plan to rearchitect the code i own so it's more agent friendly. answer open questions with prototyping rather than defer to me, and back the plan with real data. 4. before you begin, use /sharpen + /in-your-own-words and confirm this prompt with me

Ibragim@ibragim_bad·19h

Your old ~/.claude and ~/.codex sessions are an eval set: your corrections become test cases for benchmarking models, harnesses and plugins.

I had 1,270 agent sessions sitting in ~/.claude, ~/.codex and ~/.pi that’s an eval set: my corrections to the agent are the test cases turned it into the my-mini-benchmark. now I test any model, harness or plugin on my own tasks guide + the skill 👇 https://t.co/oaRtwFJTxT

ℏεsam@Hesamation·1d

A Stanford lecture on how an always-on agent works: model changes, architecture, harness design and context engineering.

Stanford just dropped a 45-min lecture from SpaceX on how Grok Bot actually works and it's basically a harness engineering masterclass: > what changed in the models to allow always-on agents > the architecture of an always-on agent > the harness design > context engineering for https://t.co/hOpUA9Iney

Daniel San@dani_avila7·1d

/export file_name dumps a Claude Code session to plain text: hand it to another agent to continue, or mine it to write custom skills.

Claude Code lets you export your entire session with /export file_name This creates a plain text file containing your session history You can use it as context for another agent to continue your work, or analyze it to create custom skills that improve your Claude Code workflow https://t.co/cMZG0J01s9

aditya@adxtyahq·1d

A 40-agents-a-day setup: Claude Code on Opus plans from the Linear ticket, then work fans out across Claude, Codex and a big VM.

i run around 40 AI agents a day, 24/7. here's how i do it. my setup: 3 Claude $200 accounts 1 Codex $100 account Grok Bot a CPU VM with 128 GB RAM and 32 cores my workflow: 1. planning: Claude Code with Opus 5.5 breaks down the Linear ticket, reasons through the architecture, https://t.co/5k7FQTqP2r

Ackerman@Yarilo7brigada·1d

Shrink CLAUDE.md to core rules and move procedures into skills that load on demand, with descriptions written as 'Use when...' triggers.

THIS IS A F*CKING GOLDEN JACKPOT Your 800-line CLAUDE.md is killing Claude. Every session starts with the model reading instructions on deploys, releases, migrations, API style and ten other things. None of that matters when the task is fix this typo. The context is clogged from the first second, and the rules that actually matter get lost in the noise. The idea: move every procedure into its own skill. A skill is a folder with a SKILL.md file, plus optional scripts and reference docs. At the start of a session, Claude only sees each skill’s name and a one-line description. It loads the full text only when the task actually calls for it. Here’s the split: - CLAUDE.md holds only what’s needed every time: stack, structure, 10–20 core rules - skills hold procedures: deploys, releases, migrations, reviews, onboarding into a module - the description decides when a skill fires, so write it as a trigger: Use when... - scripts inside a skill get executed, not read, so deterministic logic costs zero tokens The key trick is progressive loading. Twenty skills in context cost about as much as twenty lines of description. You can pack your team’s entire wiki into the project, and Claude opens the right page exactly when it’s needed. Bonus: any skill can also be invoked manually as a slash command, like /deploy or /release. The same file works both automatically and on demand. CLAUDE.md remembers the essentials. Skills know the details. Context stays clean. full tree CLAUDE.md: stack, project structure, core rules, nothing else .claude/skills/deploy/SKILL.md: deploy steps, checks, rollback .claude/skills/release/SKILL.md: versioning, changelog, tags .claude/skills/migrations/SKILL.md: how to write and verify migrations + a check script .claude/skills/code-review/SKILL.md: the team’s review checklist every skill’s description starts with - Use when... paste the tree and this prompt into Claude Code ↓ Rebuild my Claude Code setup around this tree: Read CLAUDE.md and all linked docs. Split the content into two parts: what’s needed for every task, and procedures for specific situations. Keep only the first part in CLAUDE.md. Target: under 100 lines. Move each procedure into .claude/skills/<name>/SKILL.md with name and description in the frontmatter. Start each description with “Use when...” and list concrete triggers. If a procedure has deterministic steps (checks, generation, validation), move them into scripts inside the skill folder and call them from there. Find procedures I repeat that aren’t written down anywhere, and propose them as new skills. Show me every change as a diff first. No edits until I say go.” ↳ https://t.co/wK9W7spslN

Kaiju@kaijucode·1d

A full Claude Code setup using only built-ins: CLAUDE.md, opusplan, Haiku search subagents, critic subagent, hooks, skills, worktrees, /schedule.

claude code basically runs my repo now, i just review and press enter... here's why: almost all of it is free config you set up once on day 1. here's my whole setup and what it costs: required (ships with claude code) > memory → CLAUDE.md > models → opusplan (opus plans, sonnet builds) > search → haiku subagents for every file lookup > review → a critic subagent before every merge > safety → permissions in settings.json > undo → plan mode + /rewind > context → /clear between unrelated tasks optional (add when you need it) > automation → hooks that format on every write > repeatable → skills for the prompts you retype > parallel → git worktrees > research → /deep-research > escalation → an opus advisor when you're stuck > scheduling → /schedule routines the only thing on this list you need to pay for is your claude plan. everything else is included with claude code.

Vercel@vercel·2d

Vercel's $1M sandbox-escape challenge: 1,285 reports, $242K paid, one Linux kernel zero-day, no customer data compromised.

We offered $1M to escape Vercel Sandbox. • 1,285 reports • $242,000 in bounties paid out • One Linux kernel zero day discovered • No reports of customer data compromise https://t.co/IWbHD6MNqJ

Theo - t3.gg@theo·2d

Theo flags that Codex still omits the current model name from its system prompt, so the agent can't tell which model it is.

It is so stupid that Codex STILL does not include the current model in the system prompt. Your agent has no idea what it is beyond "based on GPT-6". I've complained about this for a year straight and they won't fix it for no good reason. https://t.co/PepE5yeaZH

dex@dexhorthy·2d

Dex Horthy on idle compaction: compacting before the cache expires beats paying the cache miss when you return.

this is a good idea - always confused why claude would ask me if i wanna compact after 24h to save tokens since the cache expired…like i was gonna gonna eat the cache miss on the input to compaction anyway, right!? Anyways good ship imo https://t.co/UUHZ6YKepa

lauren@poteto·2d

pstack's most durable part is its tiny 10-20 line principle skills; expect most skills to compress into short principles as models improve.

perhaps one of the most durable parts of pstack are its engineering principles: very concise skills (~10-20 loc) that distill down good practices into its core. as models get smarter, i expect most of pstack skills to compress into short principles or workflows (a generic list https://t.co/bDIpHhE1cM

Thariq@trq212·2d

Claude Code's Thariq: the core skill with coding agents is being in the details, caring obsessively about the right code, architecture and UX.

@fnthawar and I talked about how the skill of coding agents is "being in the details" like running a company, you can't know everything but caring obsessively about the right details is what makes something great it could be code, architecture, design, UX- whatevers needed https://t.co/MgQvz4A3mD

Lydia Hallie ✨@lydiahallie·2d

Claude Code now compacts long idle sessions right before the ~1h prompt cache expires, so resuming reprocesses only the summary.

We recently added idle compaction to Claude Code! It now compacts long idle sessions right before the prompt cache expires (~1h). If you come back to it later, it only has to reprocess the summary instead of the entire conversation, which saves you some usage. https://t.co/SG3QyTNFwr

Thariq@trq212·2d

Claude Code Desktop now knows its own capabilities, like spinning off subsessions. The team calls it almost a v2.

Claude Code Desktop is now so much aware of its own capabilities and how it can work with you, like spinning off subsessions. I think of this almost like a v2, so many small details went into make it great. https://t.co/dpgj4fd8sQ

David Cramer@zeeg·2d

Sentry's David Cramer on agent-written software: junk building on junk compounds reliability problems, so the complexity battle matters more.

I'm seeing a progression of agents writing software that I have mixed feelings about. I have been fighting the complexity battle - the sloppy code that for us is really hard to maintain. Junk building on junk, that ultimate compounds the reliability challenges you'll face in

Visual Studio Code@code·2d

VS Code release adds interactive canvases for agent workflows and chat handoffs that keep context.

🚀 The latest @code release brings interactive canvases for agent workflows, chat handoffs without losing context, and new ways to edit Markdown. 🧵 Here’s what’s new: https://t.co/kOo0CVRtEm

ClaudeDevs@ClaudeDevs·2d

Claude Code desktop update for running many sessions side by side: faster session start, easier work splitting, and live visibility into each.

We've been working hard on making the Claude Code desktop app great for running lots of sessions side by side. It's now much quicker to start sessions, easier to split work across a few, and you can always see what each of them is doing. Here’s what's new: https://t.co/iF1WuEueMB

dex@dexhorthy·2d

A CISO conversation on balancing useful MCP access (tickets, email, Slack) for coding agents against security risk.

a couple months back i was on a pod with @vtahowe and @ianlivingstone chatting w/ a Billions-of-market-cap CISO about how to balance giving developers tools to make coding agents useful (think, mcps to all kinds of useful context link ticketing, email, slack etc) while avoiding https://t.co/eg907hnUBF

ClaudeDevs@ClaudeDevs·2d

Public beta: Claude Platform sends OpenTelemetry traces from the Messages API, Managed Agents and SDKs to your observability stack.

Now in public beta, Claude Platform can send OpenTelemetry traces from the Messages API, Claude Managed Agents, and the Claude SDKs to your observability stack. https://t.co/gXmi5xlkfO

Matt Pocock@mattpocockuk·2d

You don't strictly need skills if your process is great, but skills make it far easier to actually follow that process every time.

Do you need agent skills if you're following a great process? Of course not It's just that skills make it easier to follow that process

Visual Studio Code@code·2d

VS Code's Agents window drafts the PR, then Agent Merge watches reviews and CI, resolves conflicts, syncs the branch and can auto-merge.

Let VS Code Handle Your Pull Request for You? See how the Agents window creates a PR with an automatically drafted title and description, then lets you review checks and track progress in one place. Agent Merge can monitor reviews and CI, resolve conflicts, keep your branch in sync, and merge automatically—with fine-grained control over when it acts. 👉🏽 https://t.co/jdaoDa8Yz1

Daniel San@dani_avila7·3d

Anthropic's code-modernization v1.0.0 Claude Code plugin modernizes legacy code and tries to prove the new version behaves like the old.

Anthropic released code-modernization v1.0.0, a Claude Code plugin for modernizing legacy codebases Save this post if you have a legacy project worth modernizing What I find interesting is that it doesn’t just rewrite code. It tries to prove the new implementation behaves like

lauren@poteto·3d

Delete most code comments: agents copy what exists, so one stale workaround comment spreads through the codebase like a weed.

ban and delete most if not all code comments! agents love copying what already exists. a seemingly harmless comment explaining some temporary workaround is one of the worst viruses you can have in your codebase. like a weed it'll spread quickly as a tool for agents to quickly

geoff@GeoffreyHuntley·3d

A team moved all engineering from per-seat Claude/Codex subscriptions to the API with the same backlog. Huntley on the discount on intelligence.

the big discount on intelligence https://t.co/6bsVItUptl “We stopped the individual engineering subscriptions to Claude and Codex on the engineers’ Macs, and routed all engineering work through the API. The team and backlog didn’t change, and we didn’t ask the engineers to

lauren@poteto·3d

New /in-your-own-words skill in pstack makes the agent restate your goal so you confirm it understood intent before it starts.

this is now a new skill /in-your-own-words in pstack 0.15.23! extremely useful to ensure that you and your agent are aligned on the goal and that it understands your intent clearly https://t.co/3dIimxHJEt

Anthropic@AnthropicAI·3d

Anthropic went from ~2,500 to 30,000+ merged changes a day as agents took over, and had to build in-house source control because human-scale dev infra became the bottleneck.

New on our Engineering blog: Creating in-house source control. In May, we merged around 2,500 code changes per day. In September, we merged more than 30,000 per day. As a result, we became bottlenecked by developer infrastructure built for human (rather than agent) scale.

Mitchell Hashimoto@mitchellh·3d

OSC7501 (deterministic agent status for terminals) is spreading: Kitty, Grok Build and Codex support it, so tools can drop status heuristics.

OSC7501 adoption updates: Kitty implements and displays in the tab title. Grok Build shipped support and Codex merged. In less than a week I think we're nearly at a place where the entire industry can drop their heuristics and get deterministic status updates.

geoff@GeoffreyHuntley·4d

Geoffrey Huntley essay: nobody should type code by hand anymore, so here's what's left of the job.

🗞️ nobody should type code by hand anymore. here's what's left. https://t.co/MOkQkM1rmB https://t.co/PEUzSgQClo

lauren@poteto·4d

poteto never lets agents write for him (docs, posts, even most prompts) because writing is thinking.

i love writing. i never let agents write for me. whether it's docs, or tweets, or shitposts, or even prompts (okay maybe that one's harder as agents manage other agents) there's a simple reason for that: i still strongly believe that writing is thinking. it's a skill i hope to

lauren@poteto·4d

Slop isn't something that happens to you. Engineers can shape the environment so agents produce good code.

something that keeps bothering me is how some people see slop as something that is happening to them, like they're some kind of helpless bystander powerless to do anything about it excuse me but you are a software engineer, you literally learned how to make sand think and make https://t.co/zrcp7PwFWl

swyx@swyx·4d

First full-length workshop on Claude flows at AI Engineer NYC was standing room only. Latent Space will publish it.

world’s first full length workshop on claude flows @cjav_dev and Harrison Stall are kicking off @aidotengineer bright and early on a monday morning to a PACKED room, standing room only #AIEngineerNY @ricmac is on the case reporting for @latentspacepod and we’ll also publish https://t.co/8Ephg5ET4Q https://t.co/XieUGbXL2u