- AI Toolkit plugin
- v0.12.0
- 6 skills
- any language
review: code and PR review skills
Run a PR through two independent correctness passes, a type-safety check and a comment audit, and read one report where every finding has a file and a line.
/plugin marketplace add noppu-labs/ai-toolkit/plugin install review@ai-toolkit
npx skills add noppu-labs/ai-toolkit/review
opencode plugin add @noppu-labs/opencode-review
01 — What it does
What a review run produces.
- passes
Two correctness passes
pr-reviewruns correctness twice: through the harness'scode-reviewskill when there is one, otherwisecorrectness-review, and again with a subagent reading the diff by hand. A finding both passes made is labelledboth passes. - outcomes
Correctness findings end in an outcome
A correctness finding names a
path:lineand ends incrash,wrong data shown,wrong data persisted,harmlessorquestion, which separates a bug from an edge case you can live with. - refutation
Suspects stay until refuted
correctness-reviewdrops a candidate only when the code disproves it, and quotes the line that does. A candidate it can neither confirm nor refute stays in, marked uncertain, with the fact that would decide it. - types
Type findings propose a shape
type-safety-reviewrules a diff or a path against numbered rules (PHP-1toPHP-5,TS-1toTS-4,PY-1toPY-5), quotes the line and proposes the type to use instead, without editing any file. - comments
Comment verdicts without edits
comment-auditgives each added comment one ofDELETE,TRIM,MOVE,KEEP,UNSUREorWRONG. It only writes a report unless you pass--apply, which makes one commit and does not push. - authors
Findings written for the author
pr-commentsrewrites the report for whoever opened the PR: one comment per finding, with a code such asCOR-01, a severity taken from a fixed label and, when the fix fits in six lines, a GitHub suggestion block.
02 — Skills
The review skills.
- pr-reviewReviews a PR, or a stack of PRs, three ways (correctness, type safety, comments) and merges it all into one report.
- correctness-reviewHunts bugs in a diff angle by angle, and keeps every suspect the code does not disprove.
- comment-auditGoes through every comment a branch adds and says whether to keep, trim, move or delete it.
- type-safety-reviewFinds loose types in a diff: any and mixed, untyped arrays that should be DTOs, unchecked casts.
- pr-commentsTurns review findings into line-anchored PR comments, graded by severity.
- writing-commentsWhat belongs in a code comment, and what to cut.
03 — How it works
What one run looks like.
pr-reviewtakes one PR number, or several in merge order for a stack, and resolves each one's base and head withgh. Each PR in a stack is reviewed against the head of the PR below it, so its diff holds only its own changes.- It counts added plus deleted lines, minus lockfiles and binaries. With
code-reviewlisted, the count sets its level:highup to 300 lines,xhighup to 3000,maxabove. Otherwise it sets how manycorrectness-reviewsubagents run: one, three or four. - It builds a structural brief for each PR, with
investigate:briefwhen theinvestigateplugin is installed and plaingitcommands when it is not, and pastes the brief into the stage subagents' prompts. - The correctness passes, the type-safety stage and the comment audit are dispatched in the same turn, so they run in parallel where the harness allows. The orchestrator reads their reports and never opens the diff itself.
- It merges what came back into one report per PR, with Correctness, Type safety and Comments sections. Every finding keeps its
path:line, so it can be pasted as a review comment. pr-commentsturns that report into comments for the author and gives a verdict ofapprove,commentorrequest_changes. A stage missing from the report is never treated as clean.
04 — Requirements
What you need.
- The
ghCLI, able to read the pull request, forpr-review. Itsoriginremote has to be the repositoryghresolves pull requests against. - Claude Code 2.1.248 or later (verified on 2.1.289) with dynamic workflows enabled, to run the
pr-review-stagesworkflow. Without thempr-reviewfollows the same steps by hand. - Optional: the
investigateplugin for a richer brief, and access to context7 so the correctness stage can check library calls against their docs, though the review goes on without either. nodeonPATH, forpr-commentsand for the task-output hook, which matches reliably on Claude Code 2.1.176 or later.
05 — Questions
Before you install.
How does the review plugin relate to investigate?
They are separate plugins and each works alone. When investigate is installed, pr-review builds each PR's structural brief with investigate:brief. Without it, pr-review builds a lighter brief from git diff and git grep.
Does the review plugin work in OpenCode?
Yes. The same skills ship as the npm package @noppu-labs/opencode-review, which targets OpenCode V2, and you call a skill with @pr-review. There pr-review uses correctness-review in place of code-review, has no workflow, and the task-output hook is left out.
Can I install it with npx skills add?
Yes. That route installs the skills only, so the workflow is missing and pr-review runs by hand. It also leaves ${CLAUDE_PLUGIN_ROOT} unresolved, so for OpenCode the npm package is the supported route.
Which languages does it cover?
The type-safety rules exist for PHP, TypeScript and Python. The correctness review checks each diff against the pitfalls of whichever language it is written in.
Does it change my code or post to the pull request?
correctness-review and type-safety-review are read-only. pr-comments leaves posting to whoever calls it. comment-audit edits comments and docs, never code, in one local commit, and only with --apply.
What is the task-output hook for?
It is a PreToolUse hook on Bash and Read that denies a call naming a .output file when that file is a symlink, as a background subagent's transcript is. Without it, such a read stops an unattended run on a permission prompt, and only the Claude Code plugin has it.
More plugins
- investigateDeterministic structural briefs and two-pass code investigations: brief, module, and spec review skills
- laravelBackend skills, a specialist agent and path-scoped rules for Laravel 13 apps: controllers, models, DTOs, policies, testing and security.
- inertia-reactFrontend skills, a specialist agent and path-scoped rules for Inertia v3 + React 19 + TypeScript: shadcn, Tailwind, state management and Vite performance.