Agent skill
cli-design-expert
Expert CLI/TUI designer for building intuitive, user-friendly, and professional command-line interfaces. Focuses on UX patterns, help systems, progressive disclosure, and developer ergonomics.
Install this agent skill to your Project
npx add-skill https://github.com/aiskillstore/marketplace/tree/main/skills/dammianmiller/cli-design-expert
SKILL.md
CLI Design Expert
Overview
This skill provides expert guidance for designing and implementing professional CLI tools with:
- Intuitive UX: Commands that work as users expect
- Progressive Disclosure: Simple by default, powerful when needed
- Excellent Help: Self-documenting commands with rich examples
- Error Recovery: Helpful errors that guide users to success
- Professional Polish: Consistent styling, colors, and output formatting
PROACTIVE USAGE
Invoke this skill before:
- Creating new CLI commands
- Designing command structures
- Writing help text and documentation
- Implementing error messages
- Adding interactive prompts
Critical Design Principles
1. Command Structure - Follow Git's Model
# Noun-verb pattern (preferred)
uam memory query # <tool> <resource> <action>
uam worktree create # <tool> <resource> <action>
# Common structure
<tool> <command> [subcommand] [arguments] [--options]
# Examples
uam init # Simple command
uam init --interactive # With flag
uam generate --output ./out # With option value
uam memory query "search term" # With argument
uam worktree create fix-bug --base develop # Full example
2. Option Naming Conventions
# Short + Long options (always provide both for common options)
-v, --version # Version
-h, --help # Help
-o, --output <path> # Output path
-f, --force # Force/overwrite
-q, --quiet # Suppress output
-d, --debug # Debug mode
-n, --dry-run # Preview without changes
# Flags (boolean) vs Options (values)
--verbose # Flag (boolean)
--format json # Option with value
--count 10 # Option with number
# Negatable flags
--color / --no-color # Allow disabling defaults
--cache / --no-cache
3. Exit Codes
// Standard exit codes
const EXIT_SUCCESS = 0; // Success
const EXIT_ERROR = 1; // General error
const EXIT_USAGE = 2; // Invalid usage/arguments
const EXIT_CONFIG = 78; // Configuration error
const EXIT_NOINPUT = 66; // Input file not found
const EXIT_CANTCREAT = 73; // Can't create output
// Usage
process.exit(EXIT_SUCCESS);
process.exit(EXIT_ERROR);
Help System Design
1. Three Levels of Help
# Level 1: Command overview (--help on root)
$ uam --help
Universal Agent Memory - AI agent memory and workflow system
Usage: uam [command] [options]
Commands:
init Initialize a new project
generate Generate CLAUDE.md and agent files
memory Manage agent memory (short-term and long-term)
worktree Git worktree management for isolated development
Options:
-v, --version Show version number
-h, --help Show help
Run 'uam <command> --help' for more information on a command.
# Level 2: Command help (--help on command)
$ uam memory --help
Manage agent memory systems
Usage: uam memory <subcommand> [options]
Subcommands:
query Search long-term memory
store Store a new memory
status Show memory system status
start Start memory services
Examples:
uam memory query "redis caching"
uam memory store lesson "Always check network policies" --tags networking --importance 8
# Level 3: Subcommand help (detailed with examples)
$ uam memory query --help
Search long-term memory using semantic similarity
Usage: uam memory query <search-term> [options]
Arguments:
search-term Keywords to search for
Options:
-l, --limit <n> Maximum results (default: 10)
-t, --tags <tags> Filter by tags (comma-separated)
--min-score <n> Minimum similarity score (0-1, default: 0.5)
--json Output as JSON
Examples:
# Basic search
uam memory query "authentication flow"
# Search with filters
uam memory query "database" --tags postgres,migration --limit 5
# JSON output for scripting
uam memory query "API design" --json | jq '.results[0]'
2. Example-Driven Documentation
// Every command should have at least 3 examples
const command = new Command('generate')
.description('Generate CLAUDE.md and agent configuration files')
.option('-o, --output <path>', 'Output directory', '.')
.option('--dry-run', 'Preview without writing files')
.addHelpText('after', `
Examples:
# Generate with defaults
$ uam generate
# Generate to specific directory
$ uam generate --output ./docs
# Preview what would be generated
$ uam generate --dry-run
# Generate for specific platform
$ uam generate --platform factory
Common Issues:
If generation fails, ensure you have a .uam.json config file.
Run 'uam init' to create one interactively.
`);
Error Message Design
1. Helpful Error Format
// ❌ BAD - Cryptic error
throw new Error('ENOENT');
// ✅ GOOD - Helpful error with solution
console.error(`
${chalk.red('Error:')} Configuration file not found
Looking for: ${chalk.cyan('.uam.json')}
Searched in: ${chalk.dim(process.cwd())}
${chalk.yellow('How to fix:')}
Run ${chalk.cyan('uam init')} to create a configuration file.
${chalk.dim('For more help: uam init --help')}
`);
2. Error Categories
interface CLIError {
code: string;
message: string;
suggestion?: string;
docs?: string;
}
const ERROR_MESSAGES: Record<string, CLIError> = {
CONFIG_NOT_FOUND: {
code: 'CONFIG_NOT_FOUND',
message: 'Configuration file .uam.json not found',
suggestion: 'Run `uam init` to create a configuration file',
docs: 'https://github.com/DammianMiller/universal-agent-memory#configuration',
},
INVALID_CONFIG: {
code: 'INVALID_CONFIG',
message: 'Configuration file is invalid',
suggestion: 'Check the JSON syntax and required fields',
docs: 'https://github.com/DammianMiller/universal-agent-memory#configuration',
},
GIT_NOT_FOUND: {
code: 'GIT_NOT_FOUND',
message: 'Not a git repository',
suggestion: 'Initialize git with `git init` or run from a git repository',
},
};
function formatError(error: CLIError): void {
console.error(chalk.red(`\nError [${error.code}]:`), error.message);
if (error.suggestion) {
console.error(chalk.yellow('\nSuggestion:'), error.suggestion);
}
if (error.docs) {
console.error(chalk.dim('\nDocumentation:'), error.docs);
}
}
3. Validation Errors
// Show all validation errors at once
function validateConfig(config: unknown): ValidationResult {
const errors: string[] = [];
if (!config || typeof config !== 'object') {
return { valid: false, errors: ['Configuration must be an object'] };
}
const c = config as Record<string, unknown>;
if (!c.project) {
errors.push('Missing required field: project');
}
if (!c.project?.name) {
errors.push('Missing required field: project.name');
}
if (c.memory?.shortTerm?.maxEntries && typeof c.memory.shortTerm.maxEntries !== 'number') {
errors.push('Invalid type: memory.shortTerm.maxEntries must be a number');
}
return { valid: errors.length === 0, errors };
}
// Display validation errors nicely
function showValidationErrors(errors: string[]): void {
console.error(chalk.red('\nConfiguration validation failed:\n'));
errors.forEach((err, i) => {
console.error(chalk.red(` ${i + 1}.`), err);
});
console.error(chalk.dim('\nCheck .uam.json and fix the issues above.'));
}
Interactive Prompts
1. Inquirer.js Patterns
import inquirer from 'inquirer';
// Grouped questions with conditional flow
async function initInteractive(): Promise<Config> {
const answers = await inquirer.prompt([
{
type: 'input',
name: 'projectName',
message: 'Project name:',
default: basename(process.cwd()),
validate: (input) => input.length > 0 || 'Project name is required',
},
{
type: 'input',
name: 'description',
message: 'Description (optional):',
},
{
type: 'list',
name: 'platform',
message: 'Primary AI platform:',
choices: [
{ name: 'Claude Code (Desktop)', value: 'claudeCode' },
{ name: 'Factory.AI', value: 'factory' },
{ name: 'VS Code', value: 'vscode' },
{ name: 'OpenCode', value: 'opencode' },
],
},
{
type: 'confirm',
name: 'enableMemory',
message: 'Enable memory system?',
default: true,
},
{
type: 'list',
name: 'memoryBackend',
message: 'Long-term memory backend:',
choices: [
{ name: 'Qdrant (local Docker)', value: 'qdrant' },
{ name: 'Qdrant Cloud', value: 'qdrant-cloud' },
{ name: 'GitHub (JSON files)', value: 'github' },
{ name: 'None', value: 'none' },
],
when: (answers) => answers.enableMemory,
},
]);
return buildConfig(answers);
}
2. Confirmation for Destructive Actions
async function handleDestructiveAction(
action: string,
details: string,
execute: () => Promise<void>
): Promise<void> {
console.log(chalk.yellow(`\n⚠️ ${action}\n`));
console.log(chalk.dim(details));
const { confirmed } = await inquirer.prompt([{
type: 'confirm',
name: 'confirmed',
message: 'Are you sure you want to proceed?',
default: false,
}]);
if (!confirmed) {
console.log(chalk.dim('Cancelled.'));
return;
}
await execute();
}
// Usage
await handleDestructiveAction(
'Delete worktree and branch',
`This will delete:\n - Worktree: .worktrees/123-feature\n - Branch: feature/123-feature`,
async () => await deleteWorktree(id)
);
Output Formatting
1. Tables
// Simple aligned table
function printTable(headers: string[], rows: string[][]): void {
const widths = headers.map((h, i) =>
Math.max(h.length, ...rows.map(r => (r[i] || '').length))
);
// Header
console.log(headers.map((h, i) => h.padEnd(widths[i]!)).join(' '));
console.log(widths.map(w => '─'.repeat(w)).join(' '));
// Rows
for (const row of rows) {
console.log(row.map((cell, i) => (cell || '').padEnd(widths[i]!)).join(' '));
}
}
// Usage
printTable(
['ID', 'Branch', 'Status', 'Created'],
[
['1', 'feature/add-auth', 'active', '2024-01-15'],
['2', 'fix/memory-leak', 'merged', '2024-01-14'],
]
);
2. JSON Output for Scripting
interface CommandOptions {
json?: boolean;
quiet?: boolean;
}
function output<T>(data: T, options: CommandOptions): void {
if (options.json) {
console.log(JSON.stringify(data, null, 2));
return;
}
if (options.quiet) {
// Minimal output for scripting
if (Array.isArray(data)) {
data.forEach(item => console.log(item.id || item));
} else {
console.log((data as { id?: string }).id || data);
}
return;
}
// Human-readable output
prettyPrint(data);
}
3. Progress Indicators
import ora from 'ora';
// Single task
const spinner = ora('Processing...').start();
try {
await doWork();
spinner.succeed('Done!');
} catch (e) {
spinner.fail('Failed');
throw e;
}
// Multi-step with status
async function runPipeline(steps: Array<{ name: string; run: () => Promise<void> }>): Promise<void> {
for (let i = 0; i < steps.length; i++) {
const step = steps[i]!;
const prefix = chalk.dim(`[${i + 1}/${steps.length}]`);
const spinner = ora(`${prefix} ${step.name}`).start();
try {
await step.run();
spinner.succeed(`${prefix} ${step.name}`);
} catch (e) {
spinner.fail(`${prefix} ${step.name}`);
throw e;
}
}
}
Color Usage
import chalk from 'chalk';
// Semantic colors
const colors = {
// Status
success: chalk.green, // ✔ Operations completed
error: chalk.red, // ✖ Errors
warning: chalk.yellow, // ⚠ Warnings
info: chalk.cyan, // ℹ Information
// Emphasis
primary: chalk.blue, // Important values
secondary: chalk.dim, // Less important
highlight: chalk.bold, // Emphasis
// Data types
path: chalk.cyan, // File paths
command: chalk.cyan, // Commands to run
code: chalk.yellow, // Code snippets
url: chalk.underline.blue, // URLs
};
// Symbols
const symbols = {
success: chalk.green('✔'),
error: chalk.red('✖'),
warning: chalk.yellow('⚠'),
info: chalk.cyan('ℹ'),
bullet: chalk.dim('•'),
arrow: chalk.dim('→'),
};
Shell Completion
// Generate completion script
program
.command('completion')
.description('Generate shell completion script')
.argument('<shell>', 'Shell type (bash, zsh, fish)')
.action((shell: string) => {
switch (shell) {
case 'bash':
console.log(generateBashCompletion());
break;
case 'zsh':
console.log(generateZshCompletion());
break;
case 'fish':
console.log(generateFishCompletion());
break;
default:
console.error(`Unknown shell: ${shell}`);
process.exit(1);
}
});
// Example bash completion
function generateBashCompletion(): string {
return `
_uam_completions() {
local cur="\${COMP_WORDS[COMP_CWORD]}"
local prev="\${COMP_WORDS[COMP_CWORD-1]}"
case "\${prev}" in
uam)
COMPREPLY=($(compgen -W "init generate memory worktree droids" -- "\${cur}"))
;;
memory)
COMPREPLY=($(compgen -W "query store status start stop" -- "\${cur}"))
;;
worktree)
COMPREPLY=($(compgen -W "create list pr cleanup" -- "\${cur}"))
;;
esac
}
complete -F _uam_completions uam
`;
}
Review Checklist
Before releasing any CLI command:
-
--helpprovides clear, example-rich documentation - Short and long options for common flags
- Exit codes are meaningful (0 = success, non-zero = error)
- Error messages explain what went wrong AND how to fix it
- Destructive actions require confirmation (unless
--force) -
--jsonoutput available for scripting -
--quiet/--verboseoptions where appropriate - Colors are used semantically and respect
NO_COLORenv var - Progress indicators for long-running operations
- Validation shows all errors at once, not one at a time
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
perigon-backend
Perigon ASP.NET Core + EF Core + Aspire conventions
perigon-agent
Pointers for Copilot/agents to apply Perigon conventions
perigon-angular
Angular 21+ standalone/Material/signal conventions for Perigon WebApp
fastapi-mastery
Comprehensive FastAPI development skill covering REST API creation, routing, request/response handling, validation, authentication, database integration, middleware, and deployment. Use when working with FastAPI projects, building APIs, implementing CRUD operations, setting up authentication/authorization, integrating databases (SQL/NoSQL), adding middleware, handling WebSockets, or deploying FastAPI applications. Triggered by requests involving .py files with FastAPI code, API endpoint creation, Pydantic models, or FastAPI-specific features.
context7-efficient
Token-efficient library documentation fetcher using Context7 MCP with 86.8% token savings through intelligent shell pipeline filtering. Fetches code examples, API references, and best practices for JavaScript, Python, Go, Rust, and other libraries. Use when users ask about library documentation, need code examples, want API usage patterns, are learning a new framework, need syntax reference, or troubleshooting with library-specific information. Triggers include questions like "Show me React hooks", "How do I use Prisma", "What's the Next.js routing syntax", or any request for library/framework documentation.
browser-use
Browser automation using Playwright MCP. Navigate websites, fill forms, click elements, take screenshots, and extract data. Use when tasks require web browsing, form submission, web scraping, UI testing, or any browser interaction.
Didn't find tool you were looking for?