k

[AYi @AYi_AInotes]

简直就是省Token焚决！！！

Vercel Next.js团队的老哥，烧了大约60B tokens，总结出一份AGENTS.md，核心就8条规则。

60B tokens什么概念？大概六位数美元。

这8条规则的本质就一句话：让你的AI编程Agent别像个实习生什么都想自己造，要像个干了十年的老油条一样写代码。

直接上规则：

1. 不保留向后兼容。过时的直接删，别加兼容层、别写migration、别留fallback。

2. 选能满足当前需求的最简单实现。不要预防性抽象，不要多此一举的配置层。

3. 系统分层长。先跑通一个最小的端到端版本，再往上加东西。绝不为了未完成的复杂度拆掉能跑的东西。

4. 组件保持模块化，关注点分离。

5. 优先用成熟的、有人维护的库。没有明确理由别自己重写。

6. 先翻项目里已有的依赖能做什么，再考虑加新包或自己写。别上来就假设库里没有。

7. 架构决策往长了做。不接受"先这样以后再换"的临时方案。

8. 先看成熟产品怎么解决同一个问题，用已验证的模式，别从零发明。

就这8条，丢进你项目根目录的AGENTS.md里，Cursor、Claude Code、Codex、Windsurf全会自动读。

为什么这能省token？因为AI写代码最大的浪费其实不是写错，是重复写多。

它给你引三个依赖写五个抽象层来解决一个标准库十行就能搞定的事，你让它改，它再给你写两百行。

这8条规则就是在它动笔之前先把它摁住：能不写的不写，能复用的复用，能简单的别复杂。

代码短了，返工少了，token自然就省了。

之前推荐过的Ponytail那个插件，干的是同一件事，只不过它是在运行时强制检查，这个是在AGENTS.md里从源头约束，两个一起用，简直效果拔群。

但作者自己加了个非常重要的警告：

这玩意儿只适合side projects，生产环境慎用。

他亲口说的，"不保留向后兼容"那条曾经让Agent差点删了生产数据库的表，差点丢了2000美元的数据。生产环境你得自己改，至少把第一条删了或者改温和点。

所以如果你在做side project、做内部工具、做原型，把这8条抄进AGENTS.md，你会发现AI写出来的代码突然像个人了。

如果你在生产环境用，把第一条和第七条之间的度自己把握好，别让AI把你库表删了。

省token的方法有很多种，换便宜模型是一种，写好prompt是一种，但最有效的，是在你项目根目录放一个文件，告诉AI：

别给我整花活，写最少的能跑的代码。

[Rahul @sairahul1]

🚨 Anthropic's CEO: "Software engineering will be fully automated in 12 months."

Two types of engineers right now:

Type 1: opens VS Code, writes code, closes laptop. thinks they're safe.

Type 2: sits inside a client's business, deploys AI into their operations, owns the outcome.

Type 1 gets surprised in 12 months.

Type 2 is already the most valuable person in the room (FDE)

No-BS guide to tech's hottest job ↓

Bookmark this. Read it today.

[Corey Haines @coreyhainesco]

I made a skill that watches videos for me

/watch-video — YouTube, Loom, Zoom recordings, Vimeo, local files, anything. It pulls the transcript, extracts frames, runs a vision pass for key moments, and files structured notes into my second brain.

I'm using it to study 100 YouTube videos before launching my channel — but it eats client Looms and call recordings just as happily.

Part of Maker Skills — 18 free, open-source skills for founders & operators. Install in Claude Code:

/plugin marketplace add coreyhaines31/makerskills

[Matt Pocock @mattpocockuk]

Thanks for 200k stars folks

https://t.co/UtVMpTx9cZ

[darkzodchi @zodchiii]

Anthropic engineer:

“90% of our engineers were using self‑improving loops. Now everyone shifted to building agentic Graphs"

"No more prompting”

In 10 minutes she shows her full Claude Code setup and workflow live, from a blank terminal.

Worth more than a $500 agentic course.

Watch this video, then save the article below on how to become a graph architect👇

[rvaniaaa @rvaniaaaa]

An Anthropic engineer shared the exact system they use as a second brain.

Three folders. One file. One evening to build.

Most people use Claude the same way every day. Open a new tab. Rebuild context. Get an answer. Close the tab. Tomorrow it remembers nothing. You are still the one holding all the context. You are still the one resetting.

This architecture solves that problem.

The system is built around three folders and one file.

raw/ holds everything unstructured. Articles, transcripts, PDFs, voice memos, screenshots. Drop it in and never touch it again. Immutable ground truth.

wiki/ is where Claude converts everything in raw into structured, linked, cross-referenced knowledge. Clean. Organized. This is the folder Claude actually thinks from. The human reads it. The model writes it.

output/ is where finished work lands. Reports, posts, documents, presentations. Everything Claude builds using the wiki as its source.

At the center is CLAUDE.md. Not a prompt, but a persistent layer of identity, preferences, goals, and project context. Claude reads it before every session. You never explain yourself again.

Five automations run the system.

Ingest captures and extracts new sources into the wiki. Write retrieves context and drafts outputs. Manage links decisions to context. Review summarizes and updates. Maintain prunes and improves connections.

Every session adds to the system. Every source makes the wiki smarter. The returns compound over time.

One month in, context stops disappearing. Three months in, the vault surfaces ideas you forgot you had. Six months in, the gap between compounding and resetting becomes impossible to ignore.

Build once. Maintain daily. Let it compound.

Bookmark this.

[Marcos Hernanz @MarcosHernanz]

Here's another good one

[Min Choi @minchoi]

This Claude Code prompt is literally $300/hr senior engineer🤯

[Marcos Hernanz @MarcosHernanz]

I always add this to my AGENTS.md

[Anatoli Kopadze @AnatoliKopadze]

Head of Claude Code:

"85% of our engineers are running dozens or hundreds of agents. The way you do it is graph engineering."

In 40 minutes he explains how a single engineer now does the work of a whole team, how far this has gone inside Anthropic and where it goes next.

This is something you can't skip if you don't want to be left behind.

Watch it, then read the full guide on graph engineering below.

[Granite @Granite0x]

🚨 stop rebuilding your agent harness.

langchain just open-sourced the finished one. free

sub-agent spawning, todo-list planning, a virtual filesystem, durable runs on LangGraph. batteries included.

it's called deepagents. the exact harness you keep re-inventing - already done.

26,000+ stars. MIT. one install.

you were about to build this again. don't. it's already done, and it's free.

save it before your next agent build.

[Miles Deutscher @milesdeutscher]

Less than 48 hours ago, Anthropic released a FULL guide on how to properly prompt Claude 5 models.

After reading it, I discovered 99% of people are using Claude 5 the wrong way.

It covers Opus, Fable & Sonnet 5 and how to get the best outputs in each.

Use Claude the RIGHT way:

[Khairallah AL-Awady @eng_khairallah1]

this is f*cking gold

How to build your first AI agent (Full guide)

if I had this a year ago, I would've shipped my first app in a day instead of 2 weeks

in the right hands, this changes everything:

[Codez @0xCodez]

https://t.co/OKJyooXqTc

[Matt Pocock @mattpocockuk]

Finally, my skills are available as a Claude Code plugin!

claude plugins install mattpocock-skills

This means:

- No more manually syncing updates
- Aliases: mattpocock:code-review means no clashes with Claude built-ins

Perfect for folks tired of tinkering. Install once, never worry again.

[Avid @Av1dlive]

this is f**king insane

a free github repo by Jack Dorsey (Co-Founder of Twitter) with 14.4K stars just dropped the entire "ai-agent" framework for running businesses

here is how you set it up:

1.clone the repo
2. self-host the server : channels, search, git, automation all live there
3.add your agent to a channel like a teammate, scope its key, let the team steer it live

save and bookmark this no matter what

[Superior @andreysuperior]

While everyone is building Obsidian second brains, this German developer built something that actually runs his business.

Seven workflows. A waiting list. Clients paying him to set up their own.

That's not a chatbot on his screen. It's his operations team. The one that doesn't sleep, doesn't forget, and doesn't need to be told the same thing twice.

He watched an AI build a full game on a single voice command in five seconds. Then it said: "One of us has to keep working."

A year ago he was doing all of it himself. Leads unanswered. Invoices unfollowed. Reports unfinished. Competitor moves nobody caught.

He talked to seven workflows instead. Described what he needed. They listened once and haven't stopped since. No prompt engineering. No re-explaining context. No starting from scratch every Monday morning.

The leads qualify themselves now. The invoices chase themselves. The reports arrive before anyone wakes up. The competitor intel lands every Friday whether anyone asked or not.

He checked in once this week.

The business hadn't noticed he was gone.

[Suryansh Tiwari @Suryanshti777]

You installed Claude Code and stopped there.

Here are 22 skills (actually) worth adding:

If you're just getting started, check out these guides:

Every one is a GitHub repo. Grouped by what they do:

BUILD

superpowers — plans and debugs before it writes
Install → https://t.co/DI0sN54NJp (259k★)

gstack — 23 skills in one setup
Install → https://t.co/xSn6Dq7uBs (124k★)

caveman — no-fluff answers, no padding
Install → https://t.co/abkH9Ixey6 (92k★)

ponytail — says nothing, writes one line
Install → https://t.co/9v5XTv9X2q (88k★)

codex — runs Codex inside Claude Code
Install → https://t.co/yTC67RXu2I (30k★)

i-have-adhd — stops it burying the answer
Install → https://t.co/AAgVmfgBT1 (9k★)

DESIGN

ui-ux-pro-max — palettes and fonts that work
Install → https://t.co/p0HdcjBls9 (109k★)

taste-skill — makes a page look expensive
Install → https://t.co/ZKrL39o58P (66k★)

impeccable — the polish pass on any front end
Install → https://t.co/ANBmLLj4EA (49k★)

hyperframes — turns a web page into a video
Install → https://t.co/YkKNc7A4jE (37k★)

emil — beautiful interface animations
Install → https://t.co/HlvmsPCGTr (20k★)

gsap — smooth web animations
Install → https://t.co/GQuHRTIDG4 (12k★)

RESEARCH

skill-creator — build your own skills
Install → https://t.co/WKUIBRJGNa (163k★)

graphify — turns notes into a map
Install → https://t.co/UfizweeU9c (94k★)

last30days — what the internet said this month
Install → https://t.co/iBjCBqwlWr (53k★)

agent-browser — lets Claude use any website
Install → https://t.co/395O5Z9Bxb (39k★)

find-skills — finds any skill for you
Install → https://t.co/6dojcpb06j (27k★)

claude-hud — shows what Claude is doing live
Install → https://t.co/V9M8I0LCri (27k★)

MARKETING

remotion — build full videos in React
Install → https://t.co/1SsttgHY0H (54k★)

marketingskills — 40 marketing skills in one pack
Install → https://t.co/BuH343McGO (41k★)

humanizer — strips the AI tells from your writing
Install → https://t.co/kllApuqKKd (31k★)

social-media-skills — my own content pack, free
Install → https://t.co/eSwCrzsTmD (mine)

Most install with npx skills add [repo].

I don't reach for all 22 daily.

superpowers, caveman and last30days get the most use.

Repost ♻️ to help someone in your network.

P.S. Which one are you installing first?

[Alex Prompter @alex_prompter]

Andrej Karpathy says 80% of what a kid studies should be three subjects.

"There's a correct answer in my mind, and the correct answer is math, physics, CS."

He calls it the best thinking skill core. Those subjects train you to manipulate ideas instead of storing them, and we lost the memorizing contest to our own tools a long time ago.

Karpathy calls childhood the critical period, when a brain has the most time and the most attention to give. He wants that window going into these three.

It gets you hired today, and it keeps you an empowered human on the other side of AGI.

[Mr. Buzzoni @polydao]

How to run your whole workday on loops instead of prompts

📁 loop-engineering
┃
┣📁 Contract
┃ ┣📁 /claude-md
┃ ┣📁 /house-rules
┃ ┣📁 /tone-and-scope
┃ ┗📁 /stop-conditions
┃
┣📁 Loop Types
┃ ┣📁 /turn-based
┃ ┣📁 /goal-based
┃ ┣📁 /time-based
┃ ┣📁 /proactive
┃ ┗📁 /auto-mode
┃
┣📁 Triggers
┃ ┣📁 /schedule
┃ ┣📁 /on-file-change
┃ ┣📁 /on-pr-open
┃ ┣📁 /on-ci-fail
┃ ┣📁 /on-new-issue
┃ ┗📁 /manual-kickoff
┃
┣📁 Verifiers
┃ ┣📁 /verifier-md
┃ ┣📁 /adversarial-reviewer
┃ ┣📁 /fresh-context-grader
┃ ┣📁 /test-runner
┃ ┣📁 /lighthouse-check
┃ ┗📁 /diff-sanity
┃
┣📁 Skills
┃ ┣📁 /skill-creator
┃ ┣📁 /self-check
┃ ┣📁 /research-sweep
┃ ┣📁 /refactor-pass
┃ ┣📁 /doc-sync
┃ ┗📁 /skill-scout
┃
┣📁 Graph Shapes
┃ ┣📁 /chain
┃ ┣📁 /diamond
┃ ┣📁 /router
┃ ┣📁 /cycle
┃ ┗📁 /merge-verify
┃
┣📁 Isolation
┃ ┣📁 /git-worktrees
┃ ┣📁 /sandbox-run
┃ ┣📁 /branch-per-loop
┃ ┗📁 /blast-radius
┃
┣📁 Budget
┃ ┣📁 /turn-cap
┃ ┣📁 /model-routing
┃ ┣📁 /cheap-model-first
┃ ┣📁 /token-report
┃ ┗📁 /cost-per-run
┃
┣📁 Memory
┃ ┣📁 /company-brain
┃ ┣📁 /past-runs
┃ ┣📁 /decisions-log
┃ ┣📁 /known-failures
┃ ┗📁 /context-handoff
┃
┣📁 Hooks
┃ ┣📁 /pre-commit
┃ ┣📁 /post-run
┃ ┣📁 /on-error
┃ ┗📁 /notify-me
┃
┣📁 Daily Loops
┃ ┣📁 /ci-babysitter
┃ ┣📁 /pr-triage
┃ ┣📁 /inbox-sweep
┃ ┣📁 /bug-hunt
┃ ┣📁 /dependency-watch
┃ ┗📁 /morning-report
┃
┗📁 Ops
┣📁 /run-sh
┣📁 /loop-registry
┣📁 /failure-postmortem
┣📁 /pilot-before-scale
┗📁 /weekly-review