oh-my-gemini-cli (OmG)

# oh-my-gemini-cli (OmG)
[![Release](https://img.shields.io/github/v/release/Joonghyun-Lee-Frieren/oh-my-gemini-cli?display_name=tag&sort=semver)](https://github.com/Joonghyun-Lee-Frieren/oh-my-gemini-cli/releases)
[![License](https://img.shields.io/github/license/Joonghyun-Lee-Frieren/oh-my-gemini-cli)](LICENSE)
[![Stars](https://img.shields.io/github/stars/Joonghyun-Lee-Frieren/oh-my-gemini-cli?style=social)](https://github.com/Joonghyun-Lee-Frieren/oh-my-gemini-cli/stargazers)
[![Gemini Extension](https://img.shields.io/badge/Gemini-Extension-0d8a83)](https://geminicli.com/extensions/?name=Joonghyun-Lee-Frierenoh-my-gemini-cli)
[![Sponsor](https://img.shields.io/badge/Sponsor-GitHub_Sponsors-ea4aaa?logo=githubsponsors&logoColor=white)](https://github.com/sponsors/Joonghyun-Lee-Frieren)

[Landing Page](https://joonghyun-lee-frieren.github.io/oh-my-gemini-cli/) | [History](docs/history.md)

[한국어](docs/README_ko.md) | [日本語](docs/README_ja.md) | [Français](docs/README_fr.md) | [中文](docs/README_zh.md) | [Español](docs/README_es.md)

Context-engineering-powered multi-agent workflow pack for Gemini CLI.

> "Claude Code's core competitiveness isn't the Opus or Sonnet engine. It's Claude Code itself. Surprisingly, Gemini works well too when attached to Claude Code."
>
> - Jeongkyu Shin (CEO of Lablup Inc.), from a YouTube channel interview

This project started from that observation:
"What if we bring that harness model to Gemini CLI?"

OmG extends Gemini CLI from a single-session assistant into a structured, role-driven engineering workflow.

## What's New in v0.3.5

- Added dynamic team assembly entrypoints:
  - `/omg:team-assemble`
  - `$team-assemble`
- Added role-specialized teammates for non-coding and mixed workloads:
  - `omg-director` (orchestration)
  - `omg-consultant` (decision criteria and strategic analysis)
  - `omg-editor` (final output structuring and quality)
- Added approval-gated team assembly protocol:
  - Propose roster first
  - Ask explicit confirmation: "Proceed with this team? (yes/no)"
  - Run lifecycle only after approval
- Added model-allocation rule for predictable cost:
  - judgment/gates -> `gemini-3.1-pro`
  - implementation-heavy execution -> `gemini-3.1-flash`
  - broad low-risk exploration -> `gemini-3.1-flash-lite`
- Added runtime-state convention for assembled roster:
  - `.omg/state/team-assembly.md`

## At A Glance

| Item | Summary |
| --- | --- |
| Delivery model | Official Gemini CLI extension (`gemini-extension.json`) |
| Core building blocks | `agents/`, `commands/`, `skills/`, `context/` |
| Main use case | Complex implementation tasks that need plan -> execute -> review loops |
| Control surface | `/omg:*` commands + `$skill` workflows + sub-agent delegation |
| Default model strategy | Judgment/acceptance gates on `gemini-3.1-pro`, implementation-heavy work on `gemini-3.1-flash`, broad low-risk exploration on `gemini-3.1-flash-lite` |

## Why OmG

| Problem in raw single-session flow | OmG response |
| --- | --- |
| Context gets mixed across planning and execution | Role-separated agents with focused responsibilities |
| Hard to keep progress visible in long tasks | Explicit workflow stages and command-driven status checks |
| Repetitive prompt engineering for common jobs | Reusable skill templates (`$plan`, `$team`, `$research`) |
| Drift between "what was decided" and "what was changed" | Review and debugging roles inside the same orchestration loop |

## Architecture

```mermaid
flowchart TD
    U["User Request"] --> G["Gemini CLI"]
    G --> E["OmG Extension"]
    E --> C["commands/omg/*.toml"]
    E --> S["skills/*/SKILL.md"]
    E --> A["agents/*.md"]
    E --> X["context/omg-core.md"]
    C --> O["Orchestration Prompt"]
    S --> O
    A --> O
    X --> O
    O --> R["Structured Result: Plan, Execution, Validation, Next Steps"]
```

## Team Workflow

```mermaid
sequenceDiagram
    participant User
    participant Director as omg-director
    participant Planner as omg-planner
    participant Architect as omg-architect
    participant Executor as omg-executor
    participant Reviewer as omg-reviewer
    participant Debugger as omg-debugger
    participant Editor as omg-editor

    User->>Director: Request team execution
    Director->>User: Propose dynamic team + approval gate
    User->>Director: Approve roster
    Director->>Planner: Define goal and constraints
    Planner->>Architect: Validate technical direction (when needed)
    Architect-->>Planner: Design feedback and risk flags
    Planner->>Director: Hand off executable slices
    Director->>Executor: Assign implementation lane
    Executor->>Reviewer: Submit implementation result
    Reviewer-->>Executor: Accept or request fix
    Reviewer->>Debugger: Trigger on failing tests or regressions
    Debugger-->>Reviewer: Root cause + patch proposal
    Reviewer->>Director: Verification result
    Director->>Editor: Package validated output
    Editor-->>User: Final validated deliverable
```

## Dynamic Te

... (truncated)