Browse the full directory of Claude Code plugins — commands, agents, skills, MCP servers, and more.
Browse plugins →Only one of these Claude Code plugins for code review is built for the job — a dedicated diff reviewer. See which bundles review and which two miss.
Every team hits the same wall: pull requests pile up faster than anyone has attention to read them, and Claude Code plugins for code review are pitched as the fix. But the category is looser than the label suggests. Of the eight plugins collected here, exactly one is a dedicated reviewer — the mattpocock code-review skill, which diffs your branch against a fixed point and reports findings through parallel sub-agents. If you already live inside a broader agent kit, everything-claude-code folds review into a full pipeline instead. And two of the eight don't review code at all: ui-ux-pro-max critiques visual designs, and nature-skills runs simulated peer review of academic papers. Install the dedicated reviewer first; reach for the kits only if you already run them.
Here is what each listing ships and what it actually reviews. Component types come from the directory; install and star counts are shown only where the snapshot has real per-plugin numbers.
| Plugin | Components | Installs (7d) | Stars | What it reviews |
|---|---|---|---|---|
| mattpocock code-review | skills | n/a | n/a | Diffs HEAD against a fixed point |
| everything-claude-code | commands, agents, skills, hooks | 586 | 226,080 | Code review as one pipeline |
| mattpocock-skills | skills | n/a | n/a | Review among many eng workflows |
| ask-matt | skills | n/a | n/a | Routes you to the right flow |
| implement | skills | n/a | n/a | TDD ending in review + commit |
| development-discipline | skills | n/a | n/a | Engineering guardrails |
| ui-ux-pro-max | skills | 444 | 100,453 | Visual design, not code |
| nature-skills | skills | 133 | 20,042 | Academic papers, not code |
The n/a stars are not missing data. Four of these plugins come from a single source repo that hosts 21 plugins, so the repo's star count is not any one plugin's star count — treating it as such would be dishonest. Judge those four on what their skills do, not on a shared vanity number.
This one does the narrow job well: it reviews the changes between HEAD and a fixed point for coding-style conformance and specification faithfulness, running parallel sub-agents that report side by side. Use it for a pre-PR self-review — point it at your feature branch and its base commit, and read both lanes before you ask a human. It differs from everything-claude-code in scope, not ambition: it is skills-only and does exactly one thing, where the larger kit ships four component types and many jobs. If review is the only task you want automated, start at its plugin page.
ECC extends Claude Code with 548 skills, 185 commands, 134 agents, and 2 hooks, and its orchestration pipelines cover TDD, code review, and bug fixing among many other flows. Use it when you want a single install that handles review alongside testing, security, and deployment rather than a purpose-built reviewer. It is the only featured plugin that ships all four component types — including agents and hooks — where every other option here is skills-only; that breadth is the reason it draws 226,080 stars against 586 installs in the last seven days. If you are assembling a broad automation stack, its review lane fits naturally, and our guide to automating Claude Code workflows covers the surrounding setup. The kit lives at everything-claude-code.
These four listings come from the same "skills for real engineers" collection, and they are designed to compose rather than compete. Review is one station on a line that runs from idea to commit.
This is the full collection: structured, AI-guided workflows for API design, refactoring, code review, TDD, prototyping, and documentation, each running a focused session that produces an artifact. Use it when you want review wired into the same toolbox as the rest of your engineering work instead of installed on its own. It differs from the standalone code-review listing by being the superset — same author, but the whole kit rather than the single reviewing skill. See the mattpocock-skills page for the skill list.
ask-matt does not review code itself; it guides you through a structured workflow — idea grilling, prototyping, PRD creation, implementation, and code review — by routing to the appropriate skill for your situation. Use it when you are not sure which of the mattpocock skills you need and want the plugin to pick. It complements the code-review skill directly: ask-matt decides when review is the next step, code-review does the reviewing. Start at the ask-matt page.
implement builds features from PRDs or issues using test-driven development, runs typechecks and tests at agreed seams, and then walks you through code review and commit. Use it when the work is scoped and you want the review to happen at the end of a disciplined build rather than as a separate pass. It pairs with the standalone reviewer: implement produces the diff and a first review, and you can hand the same branch to code-review for a focused second look. Its implement page has the details.
This listing ships skills only and carries no enrichedSummary in the snapshot, so treat it as thinly documented for review specifically. Its parent repository is a multi-harness marketplace whose skills lean toward engineering discipline and PR handling, which is why it lands in this category. Until it has a clearer per-plugin summary, prefer the mattpocock reviewer for actual diff review and use this for the broader guardrail habits it encodes. Its plugin page is the place to check whether that has changed.
Two of the eight match the "review" keyword and nothing more. Knowing that up front saves an install you would uninstall an hour later.
ui-ux-pro-max generates brand systems and accessible UI with design tokens, palettes, and shadcn/ui components, then reviews and optimizes visual quality with AI-driven art direction. The word "review" is real, but it means design critique, not code review — it never reads a diff. With 444 installs in the last seven days and 100,453 stars, it is a capable design tool in the wrong aisle here. If that is the job you actually have, our frontend design plugins roundup is the right list; the ui-ux-pro-max page covers its design work.
nature-skills generates Nature-style scientific manuscripts, figures, and response letters and runs simulated peer review over research materials. That peer review is the keyword collision — it reviews academic writing, not source changes. With 133 installs in the last seven days and 20,042 stars, it is a real research tool, but its README is largely a recruitment notice, so it is thin on anything a code reviewer needs. Confirm the mismatch yourself on the nature-skills page.
The pick changes with the job, not the star count. For a focused diff review before you open a PR, install the mattpocock code-review skill — it is the only listing scoped to exactly that. If you are standing up a wider automation stack and want review to ride along with testing and deployment, everything-claude-code's pipeline is the pragmatic call. If you already work inside the mattpocock skills, let ask-matt route you and hand finished branches from implement into the reviewer. Skip ui-ux-pro-max and nature-skills for code review entirely; they belong to design and research work. Browse the full set on the plugins directory and try the dedicated reviewer on your next branch.
Extend Claude Code with 548 skills, 185 commands, 134 agents, and 2 hooks for automating development workflows across coding, testing, security, deployment, infrastructure, AI/ML, and more. Includes orchestration pipelines for TDD, code review, bug fixing, and multi-agent collaboration.
Generate complete brand systems and accessible UI/UX designs using design tokens, color palettes, font pairings, chart types, and shadcn/ui components across multiple frontend stacks. Create branded assets, presentations, and social media banners, then review and optimize visual quality with AI-driven art direction.
Generate Nature-style scientific manuscripts, figures, response letters, patent drafts, and presentations from research materials. Conduct multi-source literature searches, citation management, paper reading with Chinese-English side-by-side rendering, and simulated peer review.
Claude Code plugins for frontend design split three ways: generate UI, audit it, or serve design tokens. Which to install for each job.
The Claude Code testing plugins category is huge, yet few write tests—here's what each of eight actually does and which one to install first.
Compare eight claude council plugins — multi-model code review, adversarial LLM debate, and perspective councils — with 7-day install and star data.
Provides structured, AI-guided workflows for common software engineering tasks: API design, refactoring, code review, TDD, prototyping, documentation, git safety, and writing. Each skill runs a focused session to produce artifacts like PRDs, issues, ADRs, or polished articles.
Guides you through a structured workflow for software development tasks including idea grilling, prototyping, PRD creation, issue management, implementation, and code review by routing to the appropriate skill or flow for your situation.
Review code changes between HEAD and a fixed point for coding style conformance and specification faithfulness using parallel sub-agents that report side by side.
Implements features from PRDs or issues using test-driven development, running typechecks and tests at agreed seams, then guides through code review and commit.
John's personal development workflow discipline: test-first implementation, scoped verification, debugging, review handling, and skill authoring.