Initial commit

commands/improve.md

@@ -0,0 +1,47 @@
+---
+allowed-tools: [Read, Grep, Glob, TodoWrite, Task]
+description: "Apply systematic improvements to code quality, performance, and maintainability"
+wave-enabled: true
+complexity-threshold: 0.6
+---
+
+# /hc:improve - Code Improvement
+
+## Purpose
+
+Apply systematic improvements to code quality, performance, maintainability, and best practices.
+
+## Usage
+
+```bash
+/hc:improve [target] [--type quality|performance|maintainability|style] [--safe]
+```
+
+## Arguments
+
+- `target` - Files, directories, or project to improve
+- `--type` - Improvement type (quality, performance, maintainability, style)
+- `--safe` - Apply only safe, low-risk improvements
+- `--preview` - Show improvements without applying them
+
+## Execution
+
+1. Use @agent-architect to analyze code for improvement opportunities
+2. @agent-architect creates improvement plan with risk assessment and patterns
+3. **PARALLEL**: Pass findings to @agent-coder AND @agent-designer simultaneously
+4. @agent-coder implements code improvements (parallel with designer)
+5. @agent-designer implements UI improvements (parallel with coder)
+6. Apply improvements with appropriate validation
+7. Optionally use @agent-test-engineer to verify improvements
+8. Report changes and quality metrics
+
+**Wave Trigger**: Activates for improvements spanning >15 files or >5 types
+
+## Claude Code Integration
+
+- Uses Task tool to orchestrate @agent-architect → parallel(@agent-coder, @agent-designer) workflow
+- Uses Read for comprehensive code analysis
+- Uses Task for batch improvements and file operations
+- Applies TodoWrite for improvement tracking
+- Maintains safety and validation mechanisms
+- Shares improvement plans between agents via MCP memory server