code-review-trigger

Type: ANALYSIS-ONLY DAIC Modes: DISCUSS, ALIGN, IMPLEMENT, CHECK (all modes) Priority: High

Trigger Reference

This skill activates on:

• Keywords: "review code", "code review", "check for bugs", "validate code", "find issues", "analyze code", "check quality", "code quality", "review this", "check this code", "inspect code", "audit code", "review changes", "review file"
• Intent Patterns: (review|check|validate|analyze).*?(code|file|changes|this), (find|detect|look for).*?(bug|issue|problem|error), code.*?quality, (inspect|audit).*?(code|file|changes)

From: skill-rules.json - code-review-trigger configuration

Purpose

Automatically trigger the /code-review slash command when users express intent to review code, check for bugs, or validate code quality using natural language. This skill eliminates the need to remember the exact /code-review syntax.

Core Behavior

In any DAIC mode:

1. Natural Language Detection

   • Detect when user wants code review without using slash command
   • Examples: "Review this code", "Check for bugs", "Is this code good?"
   • Automatically invoke /code-review with appropriate arguments

2. Context Extraction

   • Extract what should be reviewed from user's message
   • Examples: "Review src/auth.js" → args: "src/auth.js"
   • "Check for bugs in recent changes" → args: "recent changes"
   • "Review this" (with file in context) → args: infer from context

3. Slash Command Invocation

   • Call /code-review command with extracted arguments
   • Pass through to multi-aspect parallel code review system
   • Return results to user

Natural Language Examples

Triggers this skill:

• ✓ "Review this code"
• ✓ "Check for bugs in src/"
• ✓ "Validate code quality"
• ✓ "Find issues in my changes"
• ✓ "Analyze this file for problems"
• ✓ "Code review the recent commits"
• ✓ "Inspect the authentication logic"
• ✓ "Audit the API endpoints"

Does NOT trigger (too generic):

• ✗ "Code" (just mentions code)
• ✗ "Files" (no review intent)
• ✗ "What changed?" (status check, not review)

Safety Guardrails

ANALYSIS-ONLY RULES:

• ✓ NEVER call write tools (Edit, Write, MultiEdit)
• ✓ Only invokes /code-review command which is read-only
• ✓ Safe to run in any DAIC mode
• ✓ Provides analysis and suggestions only

Invocation Rules:

• Only invoke /code-review when clear review intent detected
• Extract specific targets from natural language when possible
• Default to reviewing recent changes if no target specified
• Never auto-fix issues (code-review is analysis-only)

Implementation Notes

This skill acts as a natural language wrapper around /code-review. The actual code review logic lives in .claude/commands/code-review.md, which:

• Launches parallel code-review-expert agents
• Analyzes code quality, security, performance, etc.
• Returns comprehensive review findings

This skill simply detects the intent and routes to that command.