Skip to content
Agentmart
Submit a tool

gemini-collab

This skill should be used when the user asks to 'ask Gemini', 'Gemini review', 'let Gemini check', 'Gemini analysis', 'dual AI', or mentions 'gemini' in any…

Skill by rocky2431

IndexedMIT★ 11
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/rocky2431-gemini-collab

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

This skill should be used when the user asks to 'ask Gemini', 'Gemini review', 'let Gemini check', 'Gemini analysis', 'dual AI', or mentions 'gemini' in any collaborative context for code review, project analysis, architecture opinions, or comparative verification.

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/gemini-collab

What's inside

Skill name
gemini-collab
Description
This skill should be used when the user asks to 'ask Gemini', 'Gemini review', 'let Gemini check', 'Gemini analysis', 'dual AI', or mentions 'gemini' in any collaborative context for code review, project analysis, architecture opinions, or comparative verification.
License
MIT
Instructions
62 lines in SKILL.md · view on GitHub

Files (6)

  • SKILL.md
  • evals.json · evals/
  • collab-protocol.md · references/
  • collaboration-modes.md · references/
  • gemini-cli-reference.md · references/
  • gemini-prompts.md · references/

SKILL.md (first lines)

# Gemini Collab - Dual AI Collaboration

Use Gemini CLI as an independent sub-agent within Claude Code. Claude orchestrates, Gemini provides independent analysis, Claude synthesizes the final result. All output goes through files — zero context pollution.

## Prerequisites

- Gemini CLI installed: `npm install -g @google/gemini-cli`
- Authenticated: `gemini` should work in terminal
- Verify: `gemini --version`

## Usage

```
/gemini-collab review                # Gemini reviews current changes
/gemini-collab review <file>         # Gemini reviews specific file(s)
/gemini-collab understand            # Gemini analyzes project structure
/gemini-collab opinion <topic>       # Get Gemini's take on an architecture decision
/gemini-collab compare <topic>       # Both AIs answer independently, then synthesize
/gemini-collab free <prompt>         # Free-form prompt to Gemini
```

When the user doesn't use a subcommand but mentions Gemini in a collaborative way, infer the most appropriate mode from context.

## How to Call Gemini

```bash
SESSION_PATH=".ultra/collab/$(date +%Y%m%d-%H%M%S)-gemini-<mode>"
mkdir -p "${SESSION_PATH}"

# Standard call — output goes to stdout by default (no -o flag needed)
gemini -p "Your prompt here" --yolo > "${SESSION_PATH}/output.md" 2>"${SESSION_PATH}/error.log"

# With file context piped in
cat file.py | gemini -p "Review this code" --yolo > "${SESSION_PATH}/output.md" 2>"${SESSION_PATH}/error.log"

# Always read output with Read tool, never rely on Bash stdout
```

**Important:**
- Always redirect to file with `> output.md`
- Always use `--yolo` for auto-approve in headless mode
- Use `2>"${SESSION_PATH}/error.log"` to suppress stderr noise
- Use the Read tool to read output files
- Set Bash timeout to 300000ms for large analyses

## Error Handling

- If `gemini` not found: `npm install -g @google/gemini-cli`
- If timeout (>5min): check partial output in file
- If empty output: proceed with Claude-only analysis
- Never block the workflow on Gemini failure

## Reference Files

Read these when you need details beyond what's in this SKILL.md:

- **`references/gemini-cli-reference.md`** — READ when you need advanced Gemini CLI flags (model selection, sandbox, output format). Contains full flag reference, approval modes, and pipe patterns.
- **`references/gemini-prompts.md`** — READ when constructing Gemini prompts. Contains CLI-ready prompt templates for each collaboration mode with recommended model per mode.
- **`references/collaboration-modes.md`** — READ when you need the detailed step-by-step flow for a specific mode. Contains definitions for review/understand/opinion/compare/free modes.
- **`references/collab-protocol.md`** — READ when writing synthesis reports or managing sessions. Contains core principles, synthesis template, session management, and error handling.

From rocky2431/ultra-builder-pro, MIT.

More from rocky2431/ultra-builder-pro

Signals

Stars
11
Forks
4
Open issues
0
Last commit (this folder)
Mar 8, 2026 (7 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from rocky2431

Cite this page

gemini-collab skill. Agent Mart. https://agentmart.dev/skills/rocky2431-gemini-collab (last updated 2026-10-24)