Agent Review Command

Execute code review using the code-quality-reviewer agent in isolated context with Opus model for enhanced long context analysis.

Invoke the command: /agent-review

Inputs

From current branch:

• All changes between main and HEAD (handled by agent)

Outputs

Terminal output:

• Comprehensive code review report from code-quality-reviewer agent
• Same format as /code-review with 3-phase analysis

Agent Integration

Step 1: Invoke Code-Quality-Reviewer Agent

Use Task tool to invoke the code-quality-reviewer agent:

Task Tool Parameters:

• subagent_type: 'code-quality-reviewer'
• prompt: "Review changes on current branch"
• description: "Comprehensive code review"

The agent will:

• Validate current branch (not main)
• Get changed files and full diff
• Execute review-standard skill (all 3 phases)
• Generate structured review report

Step 2: Display Agent Report

Present the agent's review report to the user.

The report includes:

• Phase 1: Documentation Quality Review
• Phase 2: Code Quality & Reuse Review
• Phase 3: Advanced Code Quality Review
• Overall Assessment with actionable recommendations

Comparison with /code-review

Feature	/agent-review	/code-review
Execution	Isolated agent context	Main conversation
Model	Opus (long context)	Current model
Best for	Large diffs (>500 lines)	Small diffs
Context	Clean, focused	Full conversation

Both use the same review-standard skill and produce equivalent quality reviews.