dandanmarcovici

PM Second Brain — AI skill for Claude Code

AI community

A second-brain context system for PMs using Claude Code daily — folder structure, file conventions, and CLAUDE.md patterns that feed the model the right context at the right time.

How to install PM Second Brain

This entry records only its repository, not the path inside it, so there is no exact command to give. Open dandanmarcovici/pm-second-brain and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What PM Second Brain does

A second-brain context system for PMs using Claude Code daily — folder structure, file conventions, and CLAUDE.md patterns that feed the model the right context at the right time.

Alternatives in AI

Repomix — 📦 Repomix is a powerful tool that packs your entire repository into a single, AI-friendly file 22.7k ★
Context Engineering Agent Skills — Context engineering patterns for agent skills 13.9k ★
Iai Personal Memory Engine — A cyber brain for your AI 866 ★

README

pm-second-brain

A second-brain system built for PMs who use Claude Code or OpenAI Codex as a daily working environment.

The problem it solves: AI assistants reset every conversation. If you work with one all day, you're either re-explaining context at the start of each session, or you design something so you don't have to. This is that something.

This is not a note-taking system or a PKM framework. It's a context architecture: folder structure, file conventions, and agent instruction files that feed the right information to the model at the right moment.

How it works

Claude Code reads `CLAUDE.md` at the start of every session. Codex reads `AGENTS.md`, which routes it to the same shared instructions in `CLAUDE.md`. That's where your behavioral rules live, your file conventions, and your pointers to what's currently active.

The system has three layers:

**Your work, organized.** A folder map that distinguishes projects (a finish line) from areas (ongoing, no cadence) from processes (recurring) from ideas (not yet committed to). Clear enough for both you and the model to navigate without explanation.

**Your context, structured.** Every container gets a `context.md` (stable background), a `tasks.md` (NOW / NEXT / LATER), and a `log.md` (decisions and what changed). All three are created at setup, even if empty. Without the log, context.md absorbs the timeline and becomes unmaintainable.

**Your system, contextual.** A `_system/` folder holds files that shape how Claude reasons with you: your current focus, your weekly constraints, your cognitive guardrails. `CLAUDE.md` tells the model when each one is relevant, so they don't all load every session.

Folder map

work/
  ideas/          ← untested ideas, not yet committed to
  projects/       ← time-bound work with a finish line
  areas/          ← ongoing responsibilities, no end date
  processes/      ← recurring activities (monthly, quarterly, annual)
  frameworks/     ← reusable reference