Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
Claude Code plugin for Gemini reviews and task delegation
This plugin connects Claude Code to Gemini CLI so you can send code reviews, hand off tasks, and manage background jobs from slash commands. It includes commands for review, adversarial review, rescue, status, result, cancel, and setup, plus a review gate that can block stopping when Gemini finds issues.
Builders who work in Claude Code and want Gemini to review code or take delegated tasks.
You can keep your work inside Claude Code while sending review and task work to Gemini when you need a second opinion or extra help.
What it does
Code review commands
Provides `/gemini:review` for normal read-only review and `/gemini:adversarial-review` for challenge-style review.
Task delegation commands
Provides `/gemini:rescue` and `/gemini:task` to hand work to Gemini, including background runs and model selection.
Job management
Adds `/gemini:status`, `/gemini:result`, and `/gemini:cancel` to track and stop delegated jobs.
Gemini readiness setup
Provides `/gemini:setup` to check that Gemini CLI is installed and authenticated, and to toggle the review gate.
Review gate
Can run a targeted Gemini review on Claude’s response before the session stops, so issues can be caught earlier.
How to get it
- 1Add the marketplace in Claude Code
/plugin marketplace add abiswas97/gemini-plugin-cc
- 2Install the plugin
/plugin install gemini@abiswas97-gemini
- 3Reload plugins
/reload-plugins
- 4Then run
/gemini:setup
- 5If Gemini is installed but not authenticated, set up your credentials
!gcloud auth application-default login
README
Gemini plugin for Claude Code
Use Gemini from inside Claude Code for code reviews or to delegate tasks to Gemini.
Based on openai/codex-plugin-cc, adapted for the Gemini CLI. Also adds /gemini:task for direct task delegation.
What You Get
/gemini:reviewfor a normal read-only Gemini review/gemini:adversarial-reviewfor a steerable challenge review/gemini:rescue,/gemini:status,/gemini:result, and/gemini:cancelto delegate work and manage background jobs/gemini:setupto verify Gemini CLI is ready and manage the review gate/gemini:taskfor quick one-off task delegation
Requirements
- Gemini CLI installed and authenticated — install guide
- Node.js 18.18 or later
- Google API key or Application Default Credentials — create an API key or run
gcloud auth application-default login
Install
Add the marketplace in Claude Code:
/plugin marketplace add abiswas97/gemini-plugin-cc
Install the plugin:
/plugin install gemini@abiswas97-gemini
Reload plugins:
/reload-plugins
Then run:
/gemini:setup
/gemini:setup will tell you whether Gemini is ready. If Gemini is missing, refer to the Gemini CLI installation guide.
If Gemini is installed but not authenticated, set up your credentials:
!gcloud auth application-default login
Or set the GOOGLE_API_KEY environment variable with a key from AI Studio.
After install, you should see:
- the slash commands listed below
- the
gemini:gemini-rescuesubagent in/agents
Configuration
| Variable | Purpose |
|---|---|
GOOGLE_API_KEY | API key from AI Studio. Alternative to gcloud auth application-default login. |
GEMINI_ACP_INIT_TIMEOUT_MS | Override the 30s ACP initialize timeout. Bump it if Gemini CLI cold-start is slow on your machine (e.g. Homebrew installs that fall back to a file-based keychain). |
Commands
| Command | Description |
|---|---|
/gemini:review | Code review on current work or branch. Supports --wait, --background, --base <ref>. |
/gemini:adversarial-review | Challenges design choices, not just bugs. Same flags as review plus custom focus text. |
/gemini:rescue | Delegate a task to Gemini via subagent. Supports --background, --resume, --fresh, --model <pro|flash|pro-3|flash-3>. |
/gemini:task | Quick one-off task delegation. Supports --background, --model. |
/gemini:status | Show running and recent jobs. |
/gemini:result | Show output for a finished job. Includes session ID for gemini resume. |
/gemini:cancel | Cancel an active background job. |
/gemini:setup | Check Gemini CLI readiness. Toggle review gate with --enable-review-gate / --disable-review-gate. |
Typical Flows
# Review before shipping
/gemini:review --background
/gemini:status
/gemini:result
# Hand a problem to Gemini
/gemini:rescue investigate why the build is failing in CI
# Challenge your design
/gemini:adversarial-review --base main question the caching strategy
Model Selection
The --model flag accepts shortcuts or full model names:
flash→gemini-2.5-flash(default)pro→gemini-2.5-proflash-3→gemini-3-flash-previewpro-3→gemini-3-pro-preview
If omitted, defaults to gemini-2.5-flash. Any concrete model name (e.g. gemini-3.1-pro-preview) is passed through as-is.
Review Gate
When enabled (/gemini:setup --enable-review-gate), the plugin runs a targeted Gemini review on Claude's response before stopping. If issues are found, the stop is blocked so Claude can address them first.
[!WARNING] The review gate can create a long-running Claude/Gemini loop. Only enable it when actively monitoring the session.
Development
pnpm test # run tests
pnpm run ci # type check + lint + test
pnpm run lint:fix # auto-fix lint issues
Project Structure
plugins/gemini/
├── .claude-plugin/plugin.json # Plugin metadata
├── agents/ # Gemini rescue subagent
├── commands/ # Slash commands (.md)
├── hooks/ # Session lifecycle + review gate hooks
├── prompts/ # Prompt templates
├── schemas/ # Output JSON schemas (review, errors)
├── scripts/ # Companion CLI + runtime modules
└── skills/ # Claude Code skills
License
Apache 2.0 — see LICENSE.
Files in the repo
- .claude
- .claude-plugin
- .github
- plugins
- scripts
- tests
- .gitignore
- biome.json
- CHANGELOG.md
- LICENSE
- NOTICE
- package.json
- pnpm-lock.yaml
- README.md
- tsconfig.scripts.json
Discussion (0)
Ask about usage, or say what you built with itSign in to join the discussion.
No comments yet. Be the first to say what this is good for.
More plugins

Graphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more.
OmO: Just type "mass ulw" keyword with your prompt. Now you are the master of graph engineering.
Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More
Opinionated Oxlint rules for rejecting low-evidence TypeScript and JavaScript patterns
Teams-first Multi-agent orchestration for Claude Code