{
  "markdown": "# Gemini plugin for Claude Code\n\nUse Gemini CLI from inside Claude Code for code reviews, whole-codebase analysis, or to delegate tasks.\n\nThis plugin is for Claude Code users who want to leverage Gemini's **1M token context window** alongside Claude Code's coding capabilities — giving you a \"satellite view\" of your codebase plus hands-on implementation power.\n\n## Install\n\n### From the personal plugin marketplace\n\n> **Note:** This plugin is pending approval for the official Claude Code marketplace. Until then, install from the personal marketplace below.\n\n**Step 1** — Add the marketplace registry:\n```\n/plugins marketplace add arescope/claude-plugins\n```\n\n**Step 2** — Install the Gemini plugin from the registry.\n\n### From GitHub URL (direct)\n```\n/install-plugin https://github.com/arescope/gemini-plugin-cc\n```\n\n### From source (local development)\n```bash\ngit clone https://github.com/arescope/gemini-plugin-cc.git\ncd your-project\nclaude --plugin-dir /path/to/gemini-plugin-cc\n```\n\n### Prerequisites\n\n1. Install Gemini CLI: `npm install -g @google/gemini-cli`\n2. Authenticate: API key from [AI Studio](https://aistudio.google.com/apikey) or `gemini auth` for OAuth\n\nThen verify:\n```\n/gemini:setup\n```\n\n## What You Get\n\n- `/gemini:review` — code review with full codebase context\n- `/gemini:challenge` — challenge review that questions design decisions\n- `/gemini:analyze` — whole-codebase analysis (architecture, patterns, security, dependencies)\n- `/gemini:consult` — consult Gemini on a task or delegate work\n- `/gemini:setup` — check installation, auth, and manage the review gate\n- `/gemini:status`, `/gemini:result`, `/gemini:cancel` — manage background jobs\n- `gemini-consult` subagent — fires proactively for second opinions or delegation\n\n## Architecture\n\nAll commands route through a single Node.js companion entry point (`scripts/gemini-companion.mjs`). Zero external dependencies — Node.js builtins only.\n\n```\n/gemini:review --background --base main\n  → companion.mjs review --background --base main\n    → resolves review target via git.mjs\n    → loads prompt template via prompts.mjs\n    → spawns detached task-worker\n      → runs gemini CLI with structured output\n      → parses JSON against schema\n      → writes result to .gemini-tasks/state.json\n```\n\n### Module library (`scripts/lib/`)\n\n| Module | Purpose |\n|--------|---------|\n| `args.mjs` | Parse slash-command arguments |\n| `fs.mjs` | Filesystem helpers (stdin, JSON) |\n| `gemini.mjs` | Gemini CLI wrapper — alias resolution, effort mapping, output parsing |\n| `git.mjs` | Git helpers — repo check, review target, context collection |\n| `job-control.mjs` | Job lifecycle — sort, resolve, snapshot |\n| `process.mjs` | Process management — binary check, tree kill |\n| `prompts.mjs` | Template loader with `{{VAR}}` interpolation |\n| `render.mjs` | Output formatting — review results, status tables, setup reports |\n| `state.mjs` | Centralized JSON state per workspace |\n| `tracked-jobs.mjs` | Job execution with progress tracking and log files |\n| `workspace.mjs` | Resolve workspace root from cwd |\n| `types.d.ts` | TypeScript type definitions for all interfaces |\n\n### Session lifecycle\n\nThree hooks via `hooks/hooks.json`:\n- **SessionStart** — propagates session ID to state\n- **SessionEnd** — kills running jobs, cleans up session state\n- **Stop** — smart review gate (skips turns that didn't make code edits)\n\n## Requirements\n\n| Requirement | Minimum | Recommended |\n|-------------|---------|-------------|\n| **Node.js** | 18.0+ | 22 LTS |\n| **Claude Code** | Latest | Latest |\n| **Gemini CLI** | 0.36.0+ | Latest |\n| **Git** | 2.0+ | Latest |\n| **OS** | Windows, macOS, Linux | — |\n\n## Usage\n\n### Review Before Shipping\n```\n/gemini:review --base main\n```\n\n### Challenge Your Design\n```\n/gemini:challenge --base main question the retry and error handling strategy\n```\n\n### Understand the Architecture\n```\n/gemini:analyze Map the system architecture, key abstractions, and data flow\n```\n\n### Delegate a Task\n```\n/gemini:consult investigate why the build is failing in CI\n```\n\n### Background Tasks\n```\n/gemini:consult --background investigate the flaky integration test\n/gemini:status\n/gemini:result\n```\n\n### Natural Language Delegation\nJust ask Claude:\n```\nAsk Gemini to review these changes as an external peer\nHave Gemini analyze the authentication flow across the entire codebase\n```\n\n## Review Gate (Stop Hook)\n\nThe review gate auto-reviews Claude's code changes before each stop — powered by Gemini. It skips turns that didn't make code edits.\n\n```\n/gemini:setup --enable-review-gate    # Enable\n/gemini:setup --disable-review-gate   # Disable\n```\n\nConfiguration is stored in `.gemini-tasks/state.json` via the companion runtime.\n\n**Warning**: This can create a Claude/Gemini loop and may drain usage limits. Only enable when actively monitoring.\n\n## Model Aliases\n\nAll commands accept short model names via `--model`:\n\n| Alias | Model | Use case |\n|-------|-------|----------|\n| `flash`, `fast` | gemini-3-flash-preview | Fastest (default) |\n| `pro`, `deep` | gemini-3.1-pro-preview | Most capable |\n| `lite` | gemini-2.5-flash-lite | Cheapest |\n| `flash-2` | gemini-2.5-flash | Previous gen fast |\n| `pro-2` | gemini-2.5-pro | Previous gen capable |\n\n```\n/gemini:review --model pro\n/gemini:consult --model flash --effort low quick check\n```\n\n## Effort Levels\n\nEffort levels via `--effort` (reserved for future Gemini CLI support):\n\n| Level | Budget |\n|-------|--------|\n| `low`, `l` | Light reasoning |\n| `medium`, `med`, `m` | Standard |\n| `high`, `h` | Deep reasoning |\n| `xhigh`, `max` | Maximum |\n| `none`, `off` | No thinking |\n| `<number>` | Exact token count |\n\n```\n/gemini:analyze --effort high     # Deep reasoning\n/gemini:consult --effort low       # Quick pass\n```\n\n## Internal Skills\n\nThree internal skills guide agent behavior (not user-invocable):\n\n| Skill | Purpose |\n|-------|---------|\n| `gemini-cli-runtime` | Contract for calling the companion from agents |\n| `gemini-result-handling` | Rules for presenting Gemini output |\n| `gemini-prompting` | Guidance for composing XML-tagged prompts |\n\nPlus `guide` — a quick reference for all commands and options.\n\n## Testing\n\n```bash\n# Node.js unit + integration tests (76 tests)\nnode --test tests/args.test.mjs tests/gemini.test.mjs tests/git.test.mjs \\\n  tests/process.test.mjs tests/render.test.mjs tests/state.test.mjs \\\n  tests/commands.test.mjs tests/runtime.test.mjs\n\n# Structure validation (91 checks)\nbash tests/test-plugin-structure.sh\n```\n\n## License\n\nApache-2.0\n",
  "bytes": 6523,
  "sha": "f6d1320842611f270bd6a841a65704e142fd86fefa1ad6888c27da41d1ef9c18",
  "repo_slug": "arescope/gemini-plugin-cc",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/plg_arescope_gemini_plugin_cc_gemini_f238bedd/readme"
}