From gemini-peer-review
Trigger peer review validation with Gemini CLI. Use without arguments for current changes, --base <branch> for specific branch comparison, or add a question for broad technical validation.
How this command is triggered — by the user, by Claude, or both
Slash command
/gemini-peer-review:gemini-peer-reviewThis command is limited to the following tools:
The summary Claude sees in its command listing — used to decide when to auto-load this command
# Gemini Peer Review Command You have been explicitly asked to run peer review validation using Google Gemini CLI. ## Parse Arguments Check for arguments in the command: - `/gemini-peer-review` - Review current changes (will ask what to review) - `/gemini-peer-review --base <branch>` - Review against specified branch - `/gemini-peer-review --uncommitted` - Review staged/unstaged/untracked changes only - `/gemini-peer-review --commit <SHA>` - Review a specific commit - `/gemini-peer-review <question>` - Validate answer to a broad technical question ## Execute **IMPORTANT:** All peer rev...
You have been explicitly asked to run peer review validation using Google Gemini CLI.
Check for arguments in the command:
/gemini-peer-review - Review current changes (will ask what to review)/gemini-peer-review --base <branch> - Review against specified branch/gemini-peer-review --uncommitted - Review staged/unstaged/untracked changes only/gemini-peer-review --commit <SHA> - Review a specific commit/gemini-peer-review <question> - Validate answer to a broad technical questionIMPORTANT: All peer review work MUST run as a subagent to keep the main context clean.
gemini-peer-reviewer agent with the Task toolBased on the arguments:
IMPORTANT: Ask the user what type of review they want.
Use AskUserQuestion tool:
question: "What would you like to review?"
header: "Review type"
options:
- label: "Changes vs branch"
description: "Compare current changes against a base branch (will ask which branch)"
- label: "Uncommitted changes"
description: "Review staged, unstaged, and untracked changes only"
- label: "Specific commit"
description: "Review changes from a specific commit (will ask for SHA)"
multiSelect: false
Ask for the base branch:
question: "Which branch should I compare against?"
header: "Base branch"
options:
- label: "main"
description: "Compare against the main branch"
- label: "develop"
description: "Compare against the develop branch"
- label: "master"
description: "Compare against the master branch"
multiSelect: false
Then the subagent will: get the diff with git diff [branch]...HEAD and review it via Gemini prompt
The subagent will: get the diff with git diff and git diff --staged and review via Gemini prompt
Ask: "What is the commit SHA to review?"
Then the subagent will: get the diff with git show [SHA] and review via Gemini prompt
Use the specified flag directly:
--base <branch>: Get diff with git diff [branch]...HEAD--uncommitted: Get diff with git diff and git diff --staged--commit <SHA>: Get diff with git show [SHA]If git reports no changes to review, inform the user: "No changes found to review. Make sure you have uncommitted changes or specify a branch with divergent commits."
This is a design/architecture validation:
Dispatch a subagent with:
Validate Claude's response to this question using Gemini CLI.
Question: [the question from arguments]
Use gemini with focused prompt about the question.
Return findings for comparison and synthesis.
After gathering the review parameters, dispatch to the gemini-peer-reviewer agent:
Use Task tool:
subagent_type: "gemini-peer-review:gemini-peer-reviewer"
prompt: |
Run peer review validation.
Type: [code-review | design | architecture | question]
[For code-review]: Branch: [branch], Focus: [focus area if any]
[For design/arch/question]: Claude's position: [summary]
Return only the synthesized peer review result.
The agent will return a synthesized result. Present it to the user:
npx claudepluginhub slb350/gemini-peer-review --plugin gemini-peer-review/codex-peer-reviewTriggers peer review validation using OpenAI Codex CLI. Defaults to symmetric blind-pass debate on git changes. Supports branches, commits, uncommitted code, profile init, classic mode, and questions.
/reviewGet code review feedback from external LLMs (Codex, Gemini) on git changes. Supports scopes (uncommitted, staged, branch, base), LLMs (codex, gemini, all), and custom models.
/reviewPerforms on-demand code review of uncommitted changes, git diff ranges, files, or branches, producing a report on spec compliance and code quality.
/candid-reviewRuns configurable code review on current changes (staged/unstaged/branch) with harsh/constructive tone and focus areas, categorizes issues with fixes, applies selected ones, supports re-review and git commit.
/reviewDispatches the reviewer agent to review current branch code changes against code quality principles.