Skip to content

automagik-dev/genie

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 

History

4,077 Commits
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 

Repository files navigation

Genie

Wishes in, PRs out.

npm downloads stars license discord


πŸš€ 117 commits this week Β· 6 releases Β· +22.7K LoC Β· 6 contributors

Metric Value (7d)
PRs merged 142 (20/day)
Avg merge time 0.7h
SHIP rate 99%
Releases (24h) 0

Commits per day (30d, all branches)

πŸ“Š Full velocity dashboard β†’

Genie is a CLI that turns one sentence into a finished pull request. You describe what you want β€” Genie interviews you, writes a plan, spawns parallel agents in isolated worktrees, reviews the code, and opens a PR. You approve. You merge. That's it.

Get started

Paste this into Claude Code, Codex, or any AI coding agent:

Install Genie and set up this project:

curl -fsSL https://raw.githubusercontent.com/automagik-dev/genie/main/install.sh | bash
genie
/wizard

That's it. The wizard interviews you, scaffolds your project, and walks you through your first wish. Relax.

Or install manually:

curl -fsSL https://raw.githubusercontent.com/automagik-dev/genie/main/install.sh | bash

What you get

  "Add dark mode"
       |
   /brainstorm ──── Genie asks questions until the idea is concrete
       |
   /wish ────────── Turns it into a plan: scope, criteria, task groups
       |
   /work ────────── Agents spawn in parallel worktrees, each on its own branch
       |
   /review ──────── 10 critics review. Severity-tagged. Nothing ships dirty.
       |
   Pull Request ─── You approve. You merge. Ship it.

Parallel agents. Not one agent doing everything sequentially β€” multiple agents working at the same time in isolated worktrees. No conflicts. No re-explaining context.

Automated review. A 10-critic council (architecture, security, DX, performance, ops, and more) reviews every design before you see it. Severity-tagged findings. CRITICAL blocks the PR.

Overnight mode. /dream β€” queue wishes before bed. Wake up to reviewed PRs.

Kanban boards. Built-in task boards with custom pipelines and stage gates.

Postgres-backed. All state lives in PostgreSQL β€” agents, tasks, events, messages. Queryable. Durable. Real-time via LISTEN/NOTIFY.

Full observability. Events, metrics, session replay, cost tracking. See everything your genie-spawned agents do β€” OTel-derived tool, cost, and API-request rows only cover sessions launched via genie spawn / genie team create; user-initiated Claude Code sessions are captured only when they export the OTLP env themselves.

Portable context. Identity, skills, memory β€” all markdown files in your repo, git-versioned. You own everything.

Any AI provider. Claude, Codex, or any OpenAI-compatible model. Bring your own agent.

Why Genie?

Without Genie

  • Re-explain context every new chat
  • One agent, one file at a time
  • Copy-paste PR descriptions manually
  • Review AI code yourself, line by line
  • Context lost mid-conversation
  • No memory between sessions

With Genie

  • Context captured once, inherited by every agent
  • Parallel agents in isolated worktrees
  • Automated severity-gated code review
  • Queue wishes overnight, wake up to PRs
  • 10-critic council catches what you'd miss
  • Persistent memory across sessions

Skills

17 built-in skills that compose into workflows:

Skill What it does
/brainstorm Explore vague ideas with guided questions
/wish Turn an idea into a scoped plan with acceptance criteria
/work Execute a wish with parallel agents
/review Severity-gated code review (SHIP or FIX-FIRST)
/council 10-perspective architectural deliberation
/dream Batch-execute wishes overnight
/trace Investigate bugs β€” reproduce, isolate, root cause
/fix Minimal targeted bug fixes
/report Deep investigation with browser + trace
/refine Transform rough prompts into structured specs
/learn Correct agent behavior from mistakes
/docs Audit and generate documentation
/pm Full project management playbook
/brain Knowledge graph with Obsidian-compatible vaults
/genie Auto-router β€” natural language to the right skill
/genie-hacks Community patterns and real-world workflows
/wizard Guided first-run onboarding

What's new in v4

v4 is a ground-up rewrite. 700 commits. 300 files. ~19K lines.

v3 v4
State JSON files + NATS PostgreSQL + LISTEN/NOTIFY
UI CLI only Full terminal UI
Memory None Optional knowledge brain
Tasks Basic Kanban boards, templates, projects
Observability Minimal OTLP, session capture, audit trail
Review Single pass 10-critic council deliberation
Stability Best effort Advisory locks, spawn watchdog, 200+ bug fixes

Full v4 release notes β†’

Observability substrate (v0)

Structured event emission, signed subscription tokens, 4-tier retention, and an external dead-man's switch. Consumer-compat promises are published at docs/observability-contract.md; the 5-phase rollout plan at docs/observability-rollout.md.

Executor read endpoint

External consumers (e.g. the omni scope-enforcer) can query ground-truth turn state directly from genie serve. Two paths are supported:

HTTP β€” GET http://127.0.0.1:<port>/executors/:id/state returns {state, outcome, closed_at} as JSON. The default port is pgserve_port + 2; override with GENIE_EXECUTOR_READ_PORT. Returns 404 for unknown IDs, 400 for non-UUID IDs, 200 otherwise. No authz β€” executor IDs are random UUIDs and the view exposes no secrets.

Direct SQL β€” connect to genie-PG as the read-only executors_reader role and SELECT state, outcome, closed_at FROM executors_public_state WHERE id = $1. Layer login credentials on top:

CREATE ROLE omni_scope_enforcer LOGIN PASSWORD '…' IN ROLE executors_reader;

Response shape (state / outcome / closed_at) is the stable boundary contract; removing or renaming fields is a coordinated breaking change.


Docs Β· v4 Release Β· Discord Β· MIT License

You describe the problem. Genie does everything else.