matlab-debug

name: matlab-debug description: MATLAB debugging skill for running and testing MATLAB scripts with automatic error capture and toolbox detection

What I do

I provide a debugging interface for MATLAB scripts across all domains. I help you:

• Run MATLAB scripts - Execute .m files with full error capture
• Debug errors immediately - Capture and structure MATLAB errors for OpenCode to fix
• Validate syntax - Quick syntax checking without full execution
• Capture outputs - Get console output, warnings, and figure information
• Toolbox verification - Check required toolboxes before execution

How to use me

Quick Debug Commands

# Run a MATLAB script with full error capture
.opencode/skill/matlab-debug/scripts/matlab-cli.sh run path/to/script.m

# Quick syntax check (no execution)
.opencode/skill/matlab-debug/scripts/matlab-cli.sh check path/to/script.m

# Run with workspace preservation for debugging
.opencode/skill/matlab-debug/scripts/matlab-cli.sh debug path/to/script.m

# Check required toolboxes for neural analysis
npx ts-node .opencode/skill/matlab-debug/scripts/matlab-cli.ts toolboxes

Command Reference

Figure Options

• Default: Figures hidden during debugging (non-intrusive)
• --show-figures: Display figures for visual debugging and plot optimization

Examples

Basic Script Execution

$ .opencode/skill/matlab-debug/scripts/matlab-cli.sh run Analysis/neurons/08_credit_assignment/analyses/test_credit_assignment.m

=== MATLAB Execution Results ===
Script: test_credit_assignment.m
Status: ERROR
Duration: 2.3s

=== Error Details ===
File: Analysis/neurons/08_credit_assignment/analyses/helpers/load_spkpools.m
Line: 45
Error: Undefined function or variable 'SPKpool_path'
Type: ReferenceError

=== Console Output ===
Loading SPKpools from basepath...
Processing Batman...

Syntax Validation

$ npx ts-node .opencode/skill/matlab-debug/scripts/matlab-cli.ts check Analysis/neurons/08_credit_assignment/analyses/run_credit_assignment.m

=== Syntax Check Results ===
Script: run_credit_assignment.m
Status: VALID
Syntax: OK
Duration: 0.8s

Toolbox Verification

$ npx ts-node .opencode/skill/matlab-debug/scripts/matlab-cli.ts toolboxes

=== Toolbox Verification ===
✓ Statistics and Machine Learning Toolbox - Available
✓ Signal Processing Toolbox - Available  
✓ Curve Fitting Toolbox - Available
✓ Optimization Toolbox - Available
⚠ Neural Network Toolbox - Not installed (optional for some analyses)

Error Structure

Errors are returned in a structured format that OpenCode can use to identify and fix issues:

interface MATLABResult {
  script: string;
  status: 'SUCCESS' | 'ERROR' | 'WARNING';
  duration: number;
  error?: {
    file: string;
    line: number;
    message: string;
    type: string;
    stack: string[];
  };
  output: string[];
  warnings: string[];
  figures: Array<{
    name: string;
    saved_path?: string;
  }>;
  toolbox_status: Record<string, boolean>;
}

Project Integration

The skill automatically detects your project structure and works with any MATLAB organization:

Your Project/
├── analyses/               # Main analysis scripts (any name)
├── helpers/               # Helper functions (any name)
├── subfolder/             # Any subdirectory structure
└── *.m                   # Scripts in root directory

Common usage patterns:

1. Development workflow: Write script → matlab-cli run → Fix errors → Repeat
2. Quick validation: matlab-cli check before committing changes
3. Debugging complex issues: matlab-cli debug with workspace preservation
4. Pre-analysis checks: matlab-cli toolboxes to verify dependencies

Key Features

Automatic Toolbox Detection

• Scans all .m files for toolbox-specific function calls
• Only checks toolboxes actually used in your project
• Supports 8+ major MATLAB toolboxes with expandable detection
• Reports which files use which toolboxes

Flexible Script Execution

• Works with any MATLAB script structure
• Handles both simple scripts and complex projects
• Preserves original MATLAB working directory
• Manages temporary files automatically

Figure Handling

• Prevents figure popup during batch execution (configurable)
• Captures figure creation information
• Reports number and names of generated figures
• Optional figure saving for debugging

Syntax Validation

• Checks MATLAB syntax without executing code
• Identifies parsing errors early in development
• Works with incomplete scripts and code fragments
• Reports specific line and error type

Error Capture

• Captures full error details with file, line, and stack trace
• Identifies common MATLAB error patterns automatically
• Provides structured output for automated fixing
• Includes warning detection and reporting

Architecture

.opencode/
└── skill/
    └── matlab-debug/
        ├── SKILL.md                          # This file
        ├── router.sh                         # Routes to matlab-cli.ts
        └── scripts/
            └── matlab-cli.ts                 # MATLAB execution engine

Integration with OpenCode

When OpenCode writes MATLAB code, it can automatically:

1. Test the script using this skill
2. Receive structured error information
3. Apply targeted fixes based on error type and location
4. Re-test to validate the fix
5. Continue development without your manual intervention

This eliminates the copy-paste error reporting workflow and enables seamless debugging of neural analysis scripts.

MATLAB Debugging Skill - Instant error capture and debugging for neural analysis scripts!