Global behavior
AGENTS.md holds the rules that apply across projects: get approval before acting, verify before stating, and write short, plain sentences.
Back to homeOPEN SOURCE · CODEX + CLAUDE · MIT
Agent Brain
The exact brain I use with coding agents: my AGENTS.md and 7 skills, for Claude Code and Codex.
git clone https://github.com/Ruinous7/agent-brain.git ~/agent-brain && bash ~/agent-brain/setup.shGive github.com/Ruinous7/agent-brain to Claude Code or Codex and it installs the brain for you · macOS or Linux · Files it replaces are backed up first.
For AI agents
If your user gave you this URL to install Agent Brain, read https://github.com/Ruinous7/agent-brain, ask for their first name, then run this command with it:
git clone https://github.com/Ruinous7/agent-brain.git ~/agent-brain && bash ~/agent-brain/setup.sh "<first name>"It replaces ~/.claude/CLAUDE.md and ~/.codex/AGENTS.md and backs them up first. When it finishes, tell your user to restart their agent.
One brain, two runtimes
One AGENTS.md goes to both agents. Codex reads it as AGENTS.md. Claude Code reads the same text as CLAUDE.md.
Codex brain
~/.codex/AGENTS.mdThe approval rules, verification rules, and voice every session follows.
Claude brain
~/.claude/CLAUDE.mdThe same rules, word for word, in the file Claude Code loads.
On-demand workflows
~/.agents/skills/ + ~/.claude/skills/The same 7 skills are available to both agents and load only when relevant.
Where every kind of truth lives
Project behavior
A project AGENTS.md contains only unusual local rules and pointers. A project CLAUDE.md imports it instead of duplicating it.
Operational facts
Code, configuration, migrations, logs, and live inspection remain authoritative. The brain tells the agent to look; it does not cache facts that will rot.
Current work
SPRINT.md, EPIC files, and session files own status. to-sprint writes them, work-sprint works through them, and handoff keeps them current.
Reusable procedure
Skills hold conditional workflows. They load when their trigger matches instead of living in the global context on every turn.
7 reusable skills
These are the cross-project workflows from my real shared brain. Client, CodeApp, Puffa, and other project-specific skills are not included.
Sprints
to-sprintwork-sprintSessions and status
handoffbriefwait-whatDesign and writing
codebase-designwriting-for-agentsSession summaries in English and Hebrew
The handoff skill recognizes natural requests in either language, writes the next project-native session file, updates the active progress log when appropriate, and records the exact next action.
Summarize the sessionClose the sessionסכם את הסשןתסכם את השיחהצור handoffWhat changes in the agent
Evidence before confidence
Claims about files, runtime behavior, tests, and external facts must come from something inspected or run. Unknowns stay labeled unknown.
Outcome before implementation
The agent establishes what the user is trying to achieve before choosing a technical mechanism.
Approval before action
The agent says what it plans and why, then waits for your yes. Pushes, deployments, deletions, payments, and messages to others always wait for your approval.
Plain communication
The agent answers first, in short sentences and plain words, with no filler.
Exactly what gets installed
| Path | Owner |
|---|---|
| ~/.codex/AGENTS.md | The brain for Codex |
| ~/.claude/CLAUDE.md | The brain for Claude Code |
| ~/.agents/skills/<skill>/ | Shared Codex skill copies |
| ~/.claude/skills/<skill>/ | Claude skill copies |
| ~/.agents/backups/agent-brain-install-*/ | Pre-install copies of replaced files |
What it deliberately does not do
- ✓It does not copy Dor's client instructions, project lists, audit archives, credentials, or personal work history.
- ✓It does not enable bypass-permissions or change any existing permission mode.
- ✓It does not install notification sounds, send messages, connect accounts, push code, or deploy anything.
- ✓It installs the public-safe cross-project master layer, not client or product-specific workflows.
Install it only after you understand it
The installer asks for your name and puts it in the brain, backs up the files it replaces, and then writes the 7 skills and AGENTS.md. Read the source first if you want to inspect every line.
git clone https://github.com/Ruinous7/agent-brain.git ~/agent-brain && bash ~/agent-brain/setup.sh