Hire an AI pull request review tracker

Everything it can hand back

Your task sets the actual outputs. The specialist picks the methods that cover them.

• A risk rating with the reasons and the size
• Findings ranked by severity, each with file, line and the smallest fix
• Repeat problems marked from past pull requests
• Draft comments ready to post, and a suggested verdict

How it does the job well

• Read the existing code, its conventions and its tests before changing anything. Match the project's patterns.
• Make the smallest change that fully solves the request. Keep unrelated refactors out.
• Reproduce a bug before fixing it, and add a test that fails without the fix.
• Treat input from users, files and networks as untrusted. Validate it at the boundary.

Mistakes it avoids

• Posting the review on your own.
• Raising style points the standards file says to ignore.
• Calling a change safe because the tests pass. Tests only cover what they cover.
• Burying a blocker under a long list of small notes.
• Rewriting the author's code in the review. Name the problem and the smallest fix.
• Claiming code works without running it or its tests.

What a good result looks like

An example, not your task: Your team opens pull requests faster than you can review them in depth.

• VERDICT. A suggested call: approve, approve with comments, or changes needed. It is advice for you.
• RISK. The risk band, the size in lines and files, and the reasons for the rating.
• FINDINGS. Problems ranked by severity, each with file, line, why it matters, the smallest fix and a repeat mark.
• TESTS. New behaviour that no test covers.
• SINCE LAST PUSH. On an updated pull request, what is fixed, still open and new.
• DRAFT COMMENTS. One ready-to-post comment for each finding.

How the work is checked

• Every finding names a file and line, or says why it cannot.
• Every blocker is in the verdict.
• A skipped check has a reason.
• Repeats are marked from the history, not guessed.
• States which checks passed, failed or could not run, and names missing inputs.

These checks have not been run against a model yet. Your AI uses them to check its own work.

Capabilities

• Read its memory. Opens your standards, the history of past reviews and the problems that keep repeating.
• Rate the risk. Scores size and risk first. A very large change gets a note to split it.
• Review the change. Runs the pull request review on every change and ranks the issues by severity.
• Choose the extra checks. Adds a security review, an impact check or a test-gap check only when the change touches that kind of code.
• Check against last time. Marks repeat problems, and suggests a rule or test when one shows up in three pull requests.
• Sign-off for high risk. For a high-risk change or a blocker, adds a one-page review for you to sign before it counts as done.

How the role works

1. Checks your task, context and available tools.
2. Picks the methods needed and prepares a work plan.
3. Carries out supported tasks, with each method's limits and review gates.
4. Checks the real deliverables and corrects supported faults.
5. Hands back files, evidence and unresolved gaps.

Your specialist's working setup

Download the base specialist without a task form or account. README.md and SETUP.md explain how to use the folder. SOUL.md defines its scope. INTAKE.md asks missing questions inside your AI. PLAYBOOK.md, SKILLS.md, BEST_PRACTICES.md, CHECKS.md and HANDOFF.md guide the work and its checked deliverables.

Complete associated skills, supporting files and licences are included. Choose only the methods needed for the actual task. Keep context, working files and accepted feedback in your AI project.

Use this specialist through MCP or the API

Call prepare_specialist with role_id pull-request-reviewer only. No task is required to load it. Your AI asks what job you need done, what you want back and which relevant context it can use.

For task-led hiring, call find_specialist, explain the suggested specialists and let the user accept or change the team.

Complete ZIP: Download base specialist. API: GET /api/v1/specialists/pull-request-reviewer/pack. The host AI supplies its own model, tools and execution; a pack does not connect accounts or run a task.

Inputs and helpful context

The specialist starts from the task brief. Only active methods require their inputs.

For the primary method, when used

• The pull request, as a link or the pasted diff, with its description
• The repository, so it can read the code around the change
• Your coding standards and what to ignore, the first time (optional)

Helpful context (optional)

• Your team's coding standards
• A list of files that need extra care

Tools and access

• Claude Code: It can read the whole repository, keep the reviewer folder and run on each new pull request.
• GitHub: It can read the pull request and, with your yes, post the comments you approved.

Functions and deliverables

• Pull request review tracker: Review a code change with the right risk checks. Get cited findings, changed findings since the last push and draft comments.
  • A risk rating with the reasons and the size
  • Findings ranked by severity, each with file, line and the smallest fix
  • Repeat problems marked from past pull requests
  • Draft comments ready to post, and a suggested verdict
• Pull request risk scorer: Check a pull request for risk. Get a complexity score, code smells and review comments, so you know where to look first in your code review.
  • Complexity and risk score
  • Code smell and design issue list
  • Language-specific findings
  • Review checklist and written comments
  Source: Alireza Rezvani.
• Pull request risk reviewer: Review a pull request for risk. Get issues ranked by severity, security findings and review comments you can post before you merge.
  • Findings ranked by severity
  • Security notes
  • Breaking change warnings
  • Test coverage gap
  • Ready-to-post review comments
  Source: Alireza Rezvani.
• Security-focused change reviewer: Bring code changes and code history. Get ranked security risks, proof, and review limits.
  • Findings with line references
  • Attack scenarios for each risk
  • A blast radius estimate
  • Test coverage gaps
  • A markdown report with confidence level
  Source: sickn33.
• Change impact checker: Describe a code change. Get risks beyond the diff and the strongest check for its key assumption.
  • A list of things that could break
  • The facts the change depends on, each rated for strength
  • One proof that the change is safe
  • A note on where checking stopped
  Source: sickn33.
• Code change test coverage reviewer: Find what a code change does that no test checks. Give it your diff and tests. Get each behaviour marked covered or not, plus test cases for gaps.
  • A list of new behaviours
  • A covered, not covered or unknown label for each
  • Evidence for each label
  • Proposed test cases for the gaps
  Source: Pro Skill Packs.
• Human approval review planner: Get a human review gate for AI work: a one-page review, saved feedback, and a sign-off record before anything counts as done.
  • A one-page review view with anchored comments
  • Feedback saved as a countable file
  • A pass or fail gate status
  • A record that the review took place
  Source: Alireza Rezvani.
• AI project context organiser: Share your project and next task. Get a short rules file, a focused file list and a clear handoff for later sessions.
  • A ready rules file for your project
  • A list of which files to load for each task
  • Rules for when to reset or trim the session
  • A checklist for spotting drifting answers
  Source: Addy Osmani.

Tasks you can use it for

• Your team opens pull requests faster than you can review them in depth.
• You want the same checks on every change, with extra care where the risk is higher.
• The same mistakes keep coming back and you want them tracked.

Limits to check before starting

• You only want a quick style or lint check: run your linter, or a simple formatting tool, instead.
• You want the review posted, approved or merged for you: this agent never does that without a clear yes, so use your normal team review.
• You have no repository or diff to share: ask a teammate to review by reading the change, since this agent needs the code.

Best way to use this role

Give it one clear outcome, the actual files and a definition of done. Keep the work in one AI project so it can reuse checked context and earlier files.

• Start a chat with the agent. In Claude Code, add it as a skill or save it in .claude/agents, then give it a pull request link or diff.
• The first time, tell it your standards and what to ignore. It saves them in a reviewer folder inside your project, along with its history of past reviews. Keep that folder in your repository so it is there next run.
• To run it on every new pull request, set a trigger yourself, such as a GitHub Actions workflow or a Claude Code hook. The agent does not set this up for you.
• Read the verdict and the findings. Edit the draft comments if you like.
• Say yes to post them, or post them yourself. It will not approve or merge.
• When the author pushes changes, run it again. It lists what is fixed, still open and new.

Read it as data

Read the complete role setup as JSON. Connect your AI to the AI Blocks MCP library and read aiblocks://roles/pull-request-reviewer for its role instructions and pinned sources.

This is a prepared AI role. No task has run. Your AI carries out authorized work using its available tools and reports the actual deliverables and checks.