phuryn

phuryn / pm-brain

Public

PM Brain OS: The Second Brain for Product Managers, Made of Markdown

www.productcompass.pmppm-brain-os claude-code second-brain
18
4
89% credibility
Found May 21, 2026 at 24 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

PM Brain is a personal knowledge management system for product managers that uses plain text files instead of complex databases. You install a skill once on your computer, then create a 'brain' folder in any project. Inside that folder, the system organizes your strategy, decisions, stakeholder information, and working notes into a searchable structure. When you feed it interviews, meetings, or documents, it reads them, extracts insights, and stores them with clear labels showing where each piece of information came from. The system includes six commands: /ingest to add new information, /prep to prepare for meetings, /review for weekly maintenance, /ideate for problem-solving, /risk for feature planning, and /plan for objectives. Everything stays on your laptop in plain markdown files you can read and edit yourself. The system is designed to preserve institutional memory without locking you into a proprietary platform.

How It Works

1
📚 You discover PM Brain

You hear about a simple system that keeps your product knowledge organized using plain text files instead of complicated databases.

2
⚙️ You install it once

You run a one-time setup that adds the skill to your computer. It works with Claude Code, an AI assistant you already use.

3
🧠 You create a brain for your product

In your project folder, you run a simple command and the system asks you a few questions about your company and priorities. A folder structure appears with sections for your knowledge, decisions, stakeholders, and more.

4
📄 You feed it your first interview

You paste a customer interview transcript. The system reads it, pulls out the key insights, and stores them in the right places across your brain. Every insight is tagged with where it came from.

5
You use it in your daily work
📋
Before a meeting

You run /prep and get a one-page brief on your stakeholder: their open asks, last concern, suggested questions.

💡
When solving problems

You run /ideate and get 3-7 directions, each tagged with the evidence behind it from your brain.

⚠️
Before launching features

You run /risk to scan for problems across five areas and draft hypotheses for any gaps.

6
🧹 Friday comes, you run the weekly sweep

The /review command reads your entire brain, flags what's drifting, and drafts what needs your attention. Small things get fixed automatically.

✨ Your second brain works

Six weeks later, when someone asks why you killed that other option, your brain remembers. Every claim is traceable to its source. Contradictions are preserved, not hidden. Your judgment stays yours.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 24 to 18 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is pm-brain?

pm-brain is a markdown-based second brain for product managers that runs inside Claude Code. Instead of relying on vector databases or AI memory systems, it stores everything as plain markdown files in a local folder. When you start a task, Claude reads your brain first. After you finish, it writes the synthesis back. Every Friday, a maintenance sweep flags drift and drafts what needs attention. The system uses six commands: /ingest for feeding in transcripts and artifacts, /prep for stakeholder briefs before meetings, /review for the weekly sweep, /ideate for evidence-grounded directions, /risk for feature risk scans, and /plan for structured draft plans. Everything lives in your repository, version-controlled with git. No cloud, no embeddings, no black-box retrieval.

Why is it gaining traction?

The hook is simplicity and auditability. Most AI memory systems become opaque black boxes by month three. pm-brain is just markdown files you can grep, edit, and understand. The provenance tagging system is the real differentiator: every claim in your brain has a source tag (documented interview, verbal stakeholder comment, your intuition, or industry knowledge), so you always know what the system is leaning on. The eval suite with 404 of 406 checks passing gives it credibility that most personal productivity tools skip entirely. It composes with pm-skills for workflow modules like JTBD interviews and RICE scoring.

Who should use this?

Product managers tired of context scattering across Notion, Linear, Slack, and their own heads. PMs who use Claude Code and want their AI to actually remember what they learned six weeks ago. Anyone running one product with judgment-heavy work and multiple stakeholders. Not for teams managing dozens of products simultaneously, and not for PMs who prefer GUI-based tools over markdown and git.

Verdict

Try it if you already live in Claude Code and want persistent, auditable memory without a vector database. The 99.5% test coverage and thorough documentation (walkthrough, glossary, architecture docs) are impressive for 18 stars, though the credibility score of 0.899% reflects its early-stage status. The architecture is sound, the docs are honest about tradeoffs, and the validation hook catches schema violations before they compound. Worth a weekend experiment if your context management is breaking down.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.