feat: add pr-review-toolkit plugin with workflow-based review orchestration

[cblecker]

Summary

• Adds a new pr-review-toolkit plugin that replaces personal skill orchestration with a Workflow-based approach, managing agent lifecycle automatically
• Five specialized review agents (code-reviewer, silent-failure-hunter, pr-test-analyzer, comment-analyzer, type-design-analyzer) run in parallel with dynamic selection based on changed file types
• Includes defensive args parsing to handle Claude stringifying the Workflow args parameter, and uses scriptPath for clean script loading

Test plan

• claude plugin validate ./pr-review-toolkit passes
• claude plugin validate . passes
• npx markdownlint-cli2 passes on plugin markdown
• uvx skillsaw --strict passes
• Invoke /pr-review-toolkit:review-pr <test-pr-url> on a real PR
• Verify agents run in parallel and findings are presented for approval
• Verify review posting works with line-level comments

Summary by CodeRabbit

• New Features

  • Added PR Review Plugin providing comprehensive pull request analysis using parallel specialized review agents for simultaneous code quality, test coverage, failure pattern, and design assessment.

• Documentation

  • Added plugin documentation and usage instructions for the new review functionality.

[coderabbitai]

Warning

Review limit reached

@cblecker, we couldn't start this review because you've reached your PR review rate limit.

More reviews will be available in 49 minutes and 18 seconds. Learn how PR review limits work.

Your organization has used up its prepaid credits, and credit purchases are no longer available. Enable the review add-on in the billing tab to keep reviews running — you're only billed for reviews past your plan's rate limits ($0.25/file).

⌛ How to resolve this issue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based credits.

🚦 How do rate limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan refill rate.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, the refill rate gradually slows as usage increases. The highest same-day bursts are limited more strictly.

Please see our Fair Usage Limits Policy for further information.

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 02246c1c-77f9-4c7f-9976-54bd65b86aa3

📥 Commits

Reviewing files that changed from the base of the PR and between 293c340 and 4073bb7.

📒 Files selected for processing (7)

• .claude-plugin/marketplace.json
• CLAUDE.md
• README.md
• pr-review-toolkit/.claude-plugin/plugin.json
• pr-review-toolkit/README.md
• pr-review-toolkit/skills/review-pr/SKILL.md
• pr-review-toolkit/skills/review-pr/review-pr.js

📝 Walkthrough

Walkthrough

A new pr-review-toolkit Claude plugin is introduced. It registers in the marketplace manifest, adds a plugin.json declaring a github dependency, and implements a review-pr skill that orchestrates multiple specialized agents in parallel to produce severity-sorted findings, then submits a GitHub PR review.

Changes

pr-review-toolkit Plugin

Layer / File(s)	Summary
Plugin registration and documentation .claude-plugin/marketplace.json, pr-review-toolkit/.claude-plugin/plugin.json, README.md, CLAUDE.md, pr-review-toolkit/README.md	Registers pr-review-toolkit in the marketplace with strict: true, creates the plugin manifest declaring a github dependency, and adds documentation entries in the root README, CLAUDE.md, and the plugin's own README describing agents, usage, and prerequisites.
review-pr skill control flow pr-review-toolkit/skills/review-pr/SKILL.md	Defines the review-pr skill: argument contract (pr-url), tool allowlist, and the full multi-phase control flow covering PR metadata collection, workflow invocation, findings deduplication against existing comments, user approval, and conditional REQUEST_CHANGES vs COMMENT review submission.
review-pr.js: schema, agent selection, prompts, and execution pr-review-toolkit/skills/review-pr/review-pr.js	Exports meta, defines FINDING_SCHEMA for structured output, builds runtime config and context string, implements selectAgents() with file-type regex heuristics, declares per-agent prompt templates in PROMPTS, runs all chosen agents in parallel via parallel() (forcing code-reviewer to use model opus), aggregates results, and returns severity-sorted { findings, positiveObservations }.

Sequence Diagram

sequenceDiagram
  actor User
  participant SKILL.md as review-pr Skill
  participant review-pr.js as review-pr.js Workflow
  participant GitHubAPI as GitHub MCP API
  participant Agents as Parallel Review Agents

  User->>SKILL.md: /review-pr <pr-url>
  SKILL.md->>GitHubAPI: fetch PR details (head SHA, changed files)
  SKILL.md->>review-pr.js: invoke workflow (owner/repo/PR#/sha/changedFiles/isLocal)
  review-pr.js->>Agents: run code-reviewer (opus) + conditional agents in parallel
  Agents-->>review-pr.js: findings[] + positiveObservations[]
  review-pr.js-->>SKILL.md: severity-sorted findings
  SKILL.md->>GitHubAPI: check existing comments and prior user reviews
  SKILL.md->>User: present deduplicated findings for approval
  User-->>SKILL.md: approve findings
  SKILL.md->>GitHubAPI: pull_request_review_write (pending) + add_comment_to_pending_review
  SKILL.md->>GitHubAPI: submit review (REQUEST_CHANGES or COMMENT)

Loading

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~25 minutes

Possibly related PRs

• cblecker/claude-plugins#10: Introduces the github MCP server plugin that pr-review-toolkit declares as a required dependency in its plugin.json.

Poem

🐇 A toolkit hops in, fresh and bright,
With agents in parallel, reviewing right!
Code-reviewer with opus leads the way,
While silent-failure-hunter joins the fray.
Findings sorted, critical first in line —
REQUEST_CHANGES or COMMENT, everything's fine! 🌟

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The PR title accurately describes the main change: introducing a new pr-review-toolkit plugin with workflow-based review orchestration, which is the primary objective of this changeset.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feat/pr-review-plugin

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@pr-review-toolkit/README.md`:
- Around line 9-11: The usage example in the README shows the command
`/pr-review:review-pr` but the plugin is registered with the namespace
`pr-review-toolkit`, which means users won't be able to run the documented
command. Update the command namespace in the usage example from
`/pr-review:review-pr` to `/pr-review-toolkit:review-pr` to match the actual
registered plugin name and ensure the documentation is accurate and runnable.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 6009d7ea-1c54-4e34-a70e-beccd1a3d4ce

📥 Commits

Reviewing files that changed from the base of the PR and between 8c333e7 and 293c340.

📒 Files selected for processing (7)

• .claude-plugin/marketplace.json
• CLAUDE.md
• README.md
• pr-review-toolkit/.claude-plugin/plugin.json
• pr-review-toolkit/README.md
• pr-review-toolkit/skills/review-pr/SKILL.md
• pr-review-toolkit/skills/review-pr/review-pr.js