Help us improve
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
By danshapiro
Get independent code reviews from a different, larger AI model that has zero context of your conversation, providing a truly fresh perspective on code, commits, plans, or files.
npx claudepluginhub danshapiro/fresheyes --plugin fresheyesIndependent code review using a different AI model behind the curtain.
Using the same model to review its own work has blind spots. Fresh Eyes sends your code to a completely independent model with zero context from your conversation — only the repo and the scope you give it. Model diversity improves correctness, so by default the skill picks a different model family from the one invoking it.
You need at least one of the following CLIs installed:
Codex CLI (for GPT provider):
npm install -g @openai/codex
Claude Code CLI (for Claude provider):
npm install -g @anthropic-ai/claude-code
Run codex or claude to ensure the one you chose is properly set up.
In Claude Code, run:
/plugin marketplace add danshapiro/fresheyes
/plugin install fresheyes@danshapiro-fresheyes
Important: The reviewer only sees committed code. Commit your changes before invoking.
In Claude Code:
Review this with fresh eyes - Review staged changes (or last commit if nothing staged)Review commit abc1234 with fresh eyes - Review a specific commitReview the files in src/auth/ with fresh eyes - Review specific filesDo a security review of src/auth/ with fresh eyes - Scoped review (user intent passed through faithfully)Review with fresh eyes using claude - Use Claude as the reviewerReview with fresh eyes using gpt - Use GPT as the reviewerBy default, the skill picks a different model family from the one invoking it. The reviewer operates independently — it receives only the scope you give it, with no conversation context.
Claude-provider reviews show live progress through the polling command and still return the final review text when complete.
cd /path/to/your/repo
bash ~/.claude/plugins/fresheyes/scripts/install-automatic-hook.sh
If your plugin directory is versioned, use that path instead:
cd /path/to/your/repo
bash ~/.claude/plugins/fresheyes@danshapiro-fresheyes/scripts/install-automatic-hook.sh
Optional overrides:
SKIP_FRESHEYES=1 git commit to bypass.FRESHEYES_SCOPE="Review the staged changes using git diff --cached." to customize the scope.FRESHEYES_ROOT=/path/to/plugin if the hook cannot find the plugin.Automatic mode uses medium reasoning effort and blocks the commit on blocking issues.
| Environment Variable | Description | Default |
|---|---|---|
FRESHEYES_PROVIDER | Which provider to use (gpt or claude) | gpt |
FRESHEYES_MODEL | Override the model name | gpt-5.5 (gpt) / opus (claude) |
FRESHEYES_MODE | Review mode (manual or automatic) | manual |
SKIP_FRESHEYES | Set to 1 to bypass pre-commit hook | unset |
FRESHEYES_SCOPE | Custom scope for pre-commit hook | unset |
FRESHEYES_ROOT | Override plugin root path | auto-detected |
MIT
Share bugs, ideas, or general feedback.
Based on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
社群版本的 Git 提交程式碼審查工具,使用多個專門代理進行自動化審查,配備驗證機制以過濾誤報
Multi-lens code review pipeline: deep review (Claude or Codex), automated fix loop, interactive walkthrough, manual promote, external-finding injection.
Multi-agent code review for Claude Code — parallel review by Codex (GPT-5.5), Gemini 3.1 Pro, and five Claude specialist subagents (one run blind), then verified synthesis.
Symmetric two-AI peer review using OpenAI Codex CLI. Both AIs review independently in a blind pass, then debate per-issue with terminal states until convergence. Catches significantly more issues than single-pass validation.
Perform a comprehensive code review of recent changes
Agent skills for roborev code review, fix, refine, and response workflows.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claim