Agent Review Protocol
Version: 1.0
Introduced: PR #352 (Phase 6f)
Command: /agent-review
Overview
The Agent Review Protocol enables automated code reviews by Agent 3 (AAA - Analysis/Testing). This slash command initiates a comprehensive code review of a specified PR, posts the review to GitHub, and updates coordination files.
Key Features:
- Automated PR context gathering
- 10-point scoring system across 5 criteria
- GitHub PR review posting via
gh pr review - Coordination file updates for tracking
Usage
Command Syntax
/agent-review <PR_NUMBER>Or for current branch:
/agent-review currentExample
/agent-review 352Review Process
Step 1: Get PR Details
# If argument is "current" or empty
gh pr view --json number,title,url,headRefName,baseRefName,body,files
# Otherwise fetch by PR number
gh pr view $PR_NUMBER --json number,title,url,headRefName,baseRefName,body,filesStep 2: Gather Context
# List all changed files
gh pr diff $PR_NUMBER --name-only
# Get actual diff for review
gh pr diff $PR_NUMBERStep 3: Update Coordination File
Updates agent-3-current.md on the agent-coordination branch:
# Agent 3 (AAA) - Analysis/Testing
**Status:** ASSIGNED
**Assignment:** Code Review PR #$PR_NUMBER
**Assigned By:** Orchestrator
**Assigned At:** 2025-12-04 12:00 UTC
---
## Review Assignment
**PR:** #$PR_NUMBER - $PR_TITLE
**URL:** $PR_URL
**Branch:** $HEAD_BRANCH -> $BASE_BRANCH
### Files to Review
$FILE_LIST
### Review Scope
- Code quality and best practices
- Architecture alignment
- Test coverage assessment
- Security considerations
- Platform parity (iOS <-> Android)
- Documentation adequacyStep 4: Perform Review
The review evaluates code against five criteria, each worth 0-2 points.
Review Criteria
1. Code Quality (0-2 points)
| Score | Criteria |
|---|---|
| 0 | Major quality issues, unreadable code |
| 1 | Minor issues, mostly clean |
| 2 | Clean, readable, follows conventions |
Checks:
- Clean, readable code
- Proper naming conventions
- DRY principles followed
- No unused code or dead paths
2. Architecture (0-2 points)
| Score | Criteria |
|---|---|
| 0 | Violates established patterns |
| 1 | Minor deviations, acceptable trade-offs |
| 2 | Fits patterns perfectly |
Checks:
- Fits existing patterns (MVVM, Repository, etc.)
- Proper separation of concerns
- No god classes or methods
- Appropriate abstraction levels
3. Testing (0-2 points)
| Score | Criteria |
|---|---|
| 0 | No tests or broken tests |
| 1 | Some tests, gaps in coverage |
| 2 | Comprehensive, meaningful tests |
Checks:
- Adequate test coverage
- Tests are meaningful, not just for coverage
- Edge cases considered
- Mocks used appropriately
4. Security (0-2 points)
| Score | Criteria |
|---|---|
| 0 | Security vulnerabilities present |
| 1 | Minor concerns, no critical issues |
| 2 | No security issues, best practices followed |
Checks:
- No hardcoded secrets
- Input validation
- Proper auth checks
- No SQL injection, XSS, etc.
5. Platform Parity (0-2 points)
| Score | Criteria |
|---|---|
| 0 | Platform inconsistencies |
| 1 | Minor differences, acceptable |
| 2 | iOS and Android fully aligned |
Checks:
- iOS and Android aligned
- Shared KMP code used appropriately
- No platform-specific hacks
- Consistent behavior across platforms
Scoring Summary
| Score | Verdict |
|---|---|
| 9-10 | APPROVED |
| 7-8 | APPROVED with minor suggestions |
| 5-6 | CHANGES REQUESTED (non-blocking) |
| 0-4 | CHANGES REQUESTED (blocking) |
Posting Review
Approval
gh pr review $PR_NUMBER --approve --body "## Agent 3 (AAA) Code Review
### Overall Score: 9/10
#### Code Quality: 2/2
Clean implementation with proper naming...
#### Architecture: 2/2
Follows established MVVM pattern...
#### Testing: 2/2
Comprehensive test coverage...
#### Security: 2/2
No security concerns...
#### Platform Parity: 1/2
Minor: Consider aligning constant names...
### Verdict: APPROVED
---
*Automated review by Agent 3 (AAA) via coordination protocol*"Comment Only
gh pr review $PR_NUMBER --comment --body "..."Request Changes
gh pr review $PR_NUMBER --request-changes --body "..."Coordination Updates
After review, update agent-3-current.md:
**Status:** COMPLETED
**Assignment:** Code Review PR #$PR_NUMBER
**Completed At:** 2025-12-04 12:30 UTC
## Review Summary
- **Score:** 9/10
- **Verdict:** APPROVED
- **Key Findings:** Clean implementation, good test coverage, minor parity suggestionExample Output
## Agent 3 Review Assignment
PR #352 - feat: iOS Tournament Guest Archers
https://github.com/user/repo/pull/352
### Files Under Review
- AddGuestSheet.swift (+199 lines)
- TournamentScoringViewModel.swift (+180 lines)
- TournamentScoringView.swift (+25 lines)
- TournamentScoringViewModelTests.swift (+159 lines)
### Review Posted
**Score: 9/10**
**Verdict: APPROVED**
Key findings:
- Clean implementation matching Android patterns
- Good error handling with GuestError enum
- Test coverage for new functionality
- Minor: Could add integration tests for Firebase writes
View review: https://github.com/user/repo/pull/352#pullrequestreview-XXXXX
Related Documentation
- Multi-Agent Coordination System - Overall system
- Contributing Guide - PR guidelines
- Tournament Test Guide - Testing patterns
Last Updated: 2025-12-04