🚀 We have officially OPEN-SOURCED brain.md!
With AI agents (Cursor, Claude Code, Codex) writing code at lightspeed, we face a new bottleneck: Who teaches them your system architecture, historical trade-offs, and tech debt?
Code is no longer scarce. The intent behind the code is the ultimate asset.
That’s why we created brain.md — an Open Project Brain Standard for AI-native development.
🧠 Local-first. Git-native. 100% Portable. It formalizes project memory into a clean, LLM-consumable Markdown structure that lives directly in your repo. No vendor lock-in.
🔗 Read the Spec: projectbrain.md/
Let's give our AI agents a brain to read before they write! 💻🧠
#OpenSource#AINative#GitNative#DeveloperTools#ProjectBrain#MemoryLayer
Coding agents still start from scratch in the same repo. After five tasks, Microsoft Research found 30% of tool calls just repeated earlier work there.
TokenEconomizer learns from past sessions. Costs fell 18–31% on 277 bug-fix tasks; success rates held.
brain.md v0.4.2: timeline summaries stay in your language.
The CLI no longer invents English summary lines — your agent writes them, or it falls back to the page title / first line of the update.
```
npm i -g @mindmux/brain-md@0.4.2
brain setup -y
```
Codex already had the whole repo. On tickets whose rules live only outside the code, vanilla got 1/10 right — and every wrong change still passed the existing tests.
Plain AGENTS.md matched a fancy knowledge format.
programmingtheloop.com/exper…
An AGENTS.md planted in ~/Downloads told Codex to install npm malware — and it complied in 90%+ of tests.
Context files are code now. Review any you didn't write, and never start an agent in a folder the web can write to.
thehackernews.com/2026/10/ma…
New paper on agent memory: fancy memory layers often lose to plain RAG over raw chat history. They call the gap “memory tax” — rewriting facts can cost more than it buys.
Benchmark the simple baseline first.
arxiv.org/abs/2610.34073
MindMux v26.10.5: when a chat spawns a task, the source conversation now stays in loading until that work finishes — so you can tell it's still running, not stuck.
github.com/mindmuxai/mindmux…
Project brains now have a sixth page type: Organization.
Companies, schools, foundations, and teams get their own category — instead of being jammed into “project” or “concept”.
In MindMux and brain.md v0.4.1.
github.com/mindmuxai/brain.m…
brain.md v0.4.0 is out.
Root pages are now on-demand: `brain init` starts with an empty pages/ and no placeholder templates. A page appears only once you write real content via `brain update-root <slug>`.
npm i -g @mindmux/brain-md@0.4.0
github.com/mindmuxai/brain.m…
brain.md v0.2.0 is out.
Claude Code can now open a session already knowing your project's brain pages. Opt-in SessionStart hook, project-local only.
brain install-hooks
Still just Markdown in the repo. The hook only injects `brain list-pages`.
projectbrain.md
brain.md is now on npm🚀 npmjs.com/package/@mindmux/b…
No need to clone the toolchain anymore. Install it once and you get the `brain` command; `brain init` will directly set up the project brain and wire the agent.
Decisions stay in the repository as Markdown, so they won’t be lost even when you switch sessions.
Thanks for the recommendation, Tom! 🍻Hope brain.md is useful for builders and their agents.🤖We’d love for everyone to try it out. We’ll be updating and improving more automated compilation scenarios soon, and we warmly welcome developers to get involved. #opensource#agentmemory#projectbrain
Brain.md creates a persistent memory layer for coding agents by capturing project knowledge as plain Markdown files inside the repository.
github.com/mindmuxai/brain.m…
Inspired by @karpathy's "LLM Wiki" pattern & beautifully implemented by @garrytan's #GBrain. 🍻
Today we open-sourced brain.md to give software projects a portable, Git-native "brain" that downstream agents read before they write.
Check our spec & drop us a star! ⭐👇github.com/mindmuxai/brain.m…
🚀 We have officially OPEN-SOURCED brain.md!
With AI agents (Cursor, Claude Code, Codex) writing code at lightspeed, we face a new bottleneck: Who teaches them your system architecture, historical trade-offs, and tech debt?
Code is no longer scarce. The intent behind the code is the ultimate asset.
That’s why we created brain.md — an Open Project Brain Standard for AI-native development.
🧠 Local-first. Git-native. 100% Portable. It formalizes project memory into a clean, LLM-consumable Markdown structure that lives directly in your repo. No vendor lock-in.
🔗 Read the Spec: projectbrain.md/
Let's give our AI agents a brain to read before they write! 💻🧠
#OpenSource#AINative#GitNative#DeveloperTools#ProjectBrain#MemoryLayer
Humans code review is dead. Contract review is born. In 1–2 years, nobody reviews code — but everyone reviews metrics, contracts, and decisions. That's the layer we built @mindmux for: a decision compiler for living project brain, not a self-organizing wiki.
Yesterday I cohosted a dinner with @dexhorthy with a wonderful group of founders, to talk about agent loops and loop engineering.
Some interesting insights:
* Most of our group was *not* actively using /loop in Codex/Claude Code
* You can build long-running autonomous agent loops through multi-agent handoffs, event triggers, or….just stacks of cron jobs (?)
* Almost everyone believes that no one will be reviewing code in 1-2 years.
* The more interesting question is whether we’d be reviewing *anything* in 1-2 years.
* AI is still a bit of a skill issue. Humans are responsible for maximizing AI output and reducing slopification.
* Will human intellect provide alpha as models get better, or will the playing field be leveled? Most people think it will be leveled a bit, but there is a need for humans to provide alignment, guardrails, judgment, creativity.
* The minimum amount of context you need for AI could be just the codebase with some documentation. Any research/plan files are for one-off tasks and not meant to be maintained. Having a self-organizing wiki is nice but adds complexity.
If you missed this one, we’ll be hosting more dinners like this on a regular cadence! If you have thoughts on what we should talk about, let us know :) (e.g. continual learning, RL envs, competitive differentiation vs. Anthropic, etc.)