Skip to content
Back to skills

Settings

ASecurity

View and modify Megazord project configuration

  • 3 stars
  • 0 votes
  • 0 copies
  • 3 views
  • Added June 8, 2026
ai-agentsdebugginggit

Security analysis

A100/100

Scanned June 8, 2026

npx -y skills add Sh3rd3n/megazord --skill settings --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Settings?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Settings
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/sh3rd3n-settings/badge)](https://www.skillsdirectory.com/skills/sh3rd3n-settings)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: settings
description: View and modify Megazord project configuration
disable-model-invocation: false
---

# /mz:settings

View current Megazord configuration and modify individual settings. Changes take effect on the next skill invocation.

Reference `@skills/init/design-system.md` for all visual output formatting.

## Step 1: Load and Display

Display the stage banner:

```
╔═══════════════════════════════════════════════╗
║  ⚡ MEGAZORD ► SETTINGS                       ║
╚═══════════════════════════════════════════════╝
```

Read `.planning/megazord.config.json`. If the file does not exist, display:

```
╔═══════════════════════════════════════════════╗
║  ✗ No Megazord config found                   ║
╠═══════════════════════════════════════════════╣
║  Run /mz:init to set up your project first.   ║
╚═══════════════════════════════════════════════╝
```

Then exit. Do not proceed further.

If the config file exists, parse it and display the current settings:

```
╔═══════════════════════════════════════════════════════╗
║  Current Settings                                     ║
╠═══════════════════════════════════════════════════════╣
║  ▸ Project                                            ║
║    Name:            {project_name}                    ║
║    Mode:            {yolo|interactive}                 ║
║    Depth:           {comprehensive|standard|quick}     ║
║    Parallelization: {✓/✗}                             ║
║    Commit docs:     {✓/✗}                             ║
║                                                       ║
║  ▸ Model                                              ║
║    Profile:         {quality|balanced|budget}          ║
║                                                       ║
║  ▸ Quality                                            ║
║    TDD:             {✓/✗}                             ║
║    Review:          {auto|manual|off}                  ║
║    Brainstorming:   {✓/✗}                             ║
║    CORTEX:          {✓/✗}                             ║
║    Debug:           {systematic|quick}                 ║
║                                                       ║
║  ▸ Workflow                                           ║
║    Research:        {✓/✗}                             ║
║    Plan check:      {✓/✗}                             ║
║    Verifier:        {✓/✗}                             ║
╚═══════════════════════════════════════════════════════╝
```

Use `✓` for `true` values and `✗` for `false` values. Display enum values as-is (e.g., "auto", "systematic").

## Step 2: Modification

Ask the user what they would like to change.

Use AskUserQuestion:
- header: "Setting" (7 chars)
- question: "What would you like to change?"
- options:
  - "Quality" -- description: "TDD, review, brainstorming, CORTEX, debug mode"
  - "Workflow" -- description: "Research, plan check, verifier agents"
  - "Model" -- description: "AI model profile (quality/balanced/budget)"
  - "Execution" -- description: "Mode, depth, parallelization, commit docs"
  - "Done" -- description: "No changes -- exit"

**If "Done":** Display "Settings unchanged." and exit.

### Quality Settings

If the user selects "Quality", present each toggle:

- header: "TDD" (3 chars), question: "Test-driven development (currently: {current value})", options: "On" / "Off"
- header: "Review" (6 chars), question: "Code review mode (currently: {current value})", options: "Auto" / "Manual" / "Off"
- header: "Brainstorm" (10 chars), question: "Socratic brainstorming (currently: {current value})", options: "On" / "Off"
- header: "CORTEX" (6 chars), question: "CORTEX adaptive thinking (currently: {current value})", options: "On" / "Off"
- header: "Debug" (5 chars), question: "Debugging approach (currently: {current value})", options: "Systematic" / "Quick"

### Workflow Settings

If the user selects "Workflow", present each toggle:

- header: "Research" (8 chars), question: "Research agent before planning (currently: {current value})", options: "On" / "Off"
- header: "Plan Check" (10 chars), question: "Plan verification agent (currently: {current value})", options: "On" / "Off"
- header: "Verifier" (8 chars), question: "Post-phase verifier agent (currently: {current value})", options: "On" / "Off"

### Model Profile

If the user selects "Model", present:

- header: "Model" (5 chars), question: "AI model profile (currently: {current value})", options:
  - "Quality" -- description: "Opus everywhere -- highest quality, higher cost"
  - "Balanced" -- description: "Opus for planning, Sonnet for execution"
  - "Budget" -- description: "Sonnet/Haiku -- fastest, lowest cost"

### Execution Settings

If the user selects "Execution", present each setting:

- header: "Mode" (4 chars), question: "Execution mode (currently: {current value})", options: "YOLO" / "Interactive"
- header: "Depth" (5 chars), question: "Planning depth (currently: {current value})", options: "Comprehensive" / "Standard" / "Quick"
- header: "Parallel" (8 chars), question: "Parallel plan execution (currently: {current value})", options: "On" / "Off"
- header: "Git" (3 chars), question: "Commit planning docs (currently: {current value})", options: "Yes" / "No"

### Write Updated Config

After collecting changes, write the updated config back to `.planning/megazord.config.json` with 2-space indentation.

## Step 3: Confirmation

Show what changed:

```
╔═══════════════════════════════════════════════╗
║  Settings Updated                             ║
╠═══════════════════════════════════════════════╣
║  {setting}: {old value} → {new value}         ║
║  {setting}: {old value} → {new value}         ║
╚═══════════════════════════════════════════════╝
```

If no values actually changed (user selected same values), display: "No changes detected. Settings remain the same."

Note: "Changes take effect on the next skill invocation."

Ask if the user wants to change anything else. If yes, return to Step 2. If no, exit with the next-up block:

```
═══════════════════════════════════════════════════
▸ Next Up
**Check Project Status** -- see current progress and next actions
`/mz:status`
═══════════════════════════════════════════════════
```

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.