mirror of
https://github.com/catlog22/Claude-Code-Workflow.git
synced 2026-02-05 01:50:27 +08:00
feat(cli): add CLI prompt simulation and testing scripts
- Introduced `simulate-cli-prompt.js` to simulate various prompt formats and display the final content passed to the CLI. - Added `test-shell-prompt.js` to test actual shell execution of different prompt formats, demonstrating correct vs incorrect multi-line prompt handling. - Created comprehensive tests in `cli-prompt-parsing.test.ts` to validate prompt parsing, including single-line, multi-line, special characters, and template concatenation. - Implemented edge case handling for empty lines, long prompts, and Unicode characters.
This commit is contained in:
catlog22
214
ccw/scripts/simulate-cli-prompt.js
Normal file
214
ccw/scripts/simulate-cli-prompt.js
Normal file
@@ -0,0 +1,214 @@
|
||||
#!/usr/bin/env node
|
||||
/**
|
||||
* CLI Prompt Simulation Script
|
||||
*
|
||||
* Simulates different prompt formats and outputs the final content passed to CLI.
|
||||
* Usage: node ccw/scripts/simulate-cli-prompt.js
|
||||
*/
|
||||
|
||||
import chalk from 'chalk';
|
||||
|
||||
// Test cases for different prompt formats
|
||||
const testCases = [
|
||||
{
|
||||
name: 'Single-line prompt',
|
||||
input: {
|
||||
prompt: 'Analyze the authentication module for security issues',
|
||||
tool: 'gemini',
|
||||
mode: 'analysis',
|
||||
},
|
||||
},
|
||||
{
|
||||
name: 'Single-line with quotes',
|
||||
input: {
|
||||
prompt: 'Fix the error: "Cannot read property \'id\' of undefined"',
|
||||
tool: 'gemini',
|
||||
mode: 'analysis',
|
||||
},
|
||||
},
|
||||
{
|
||||
name: 'Multi-line structured prompt',
|
||||
input: {
|
||||
prompt: `PURPOSE: Identify security vulnerabilities
|
||||
TASK: • Scan injection flaws • Check auth bypass
|
||||
MODE: analysis
|
||||
CONTEXT: @src/auth/**/*
|
||||
EXPECTED: Security report`,
|
||||
tool: 'gemini',
|
||||
mode: 'analysis',
|
||||
},
|
||||
},
|
||||
{
|
||||
name: '@ patterns with glob wildcards',
|
||||
input: {
|
||||
prompt: `CONTEXT: @src/**/*.{ts,tsx} @!node_modules/** @!dist/**
|
||||
TASK: Analyze TypeScript files`,
|
||||
tool: 'gemini',
|
||||
mode: 'analysis',
|
||||
},
|
||||
},
|
||||
{
|
||||
name: '@ patterns with Memory section',
|
||||
input: {
|
||||
prompt: `PURPOSE: Security audit
|
||||
CONTEXT: @src/auth/**/* @src/middleware/auth.ts | Memory: Using bcrypt for passwords, JWT for sessions
|
||||
EXPECTED: Vulnerability report`,
|
||||
tool: 'gemini',
|
||||
mode: 'analysis',
|
||||
},
|
||||
},
|
||||
{
|
||||
name: 'Full template format (from cli-tools-usage.md)',
|
||||
input: {
|
||||
prompt: `PURPOSE: Identify OWASP Top 10 vulnerabilities in authentication module to pass security audit; success = all critical/high issues documented with remediation
|
||||
TASK: • Scan for injection flaws (SQL, command, LDAP) • Check authentication bypass vectors • Evaluate session management • Assess sensitive data exposure
|
||||
MODE: analysis
|
||||
CONTEXT: @src/auth/**/* @src/middleware/auth.ts | Memory: Using bcrypt for passwords, JWT for sessions
|
||||
EXPECTED: Security report with: severity matrix, file:line references, CVE mappings where applicable, remediation code snippets prioritized by risk
|
||||
CONSTRAINTS: Focus on authentication | Ignore test files`,
|
||||
tool: 'gemini',
|
||||
mode: 'analysis',
|
||||
},
|
||||
},
|
||||
{
|
||||
name: 'Special characters and Unicode',
|
||||
input: {
|
||||
prompt: `TASK: • 分析代码 • Check → errors • Fix ✗ issues
|
||||
EXPECTED: Report with ✓ checkmarks`,
|
||||
tool: 'gemini',
|
||||
mode: 'analysis',
|
||||
},
|
||||
},
|
||||
{
|
||||
name: 'Code-like content',
|
||||
input: {
|
||||
prompt: `Fix: const result = arr.filter(x => x > 0).map(x => x * 2);
|
||||
Error at line 42: TypeError: Cannot read property 'length' of null`,
|
||||
tool: 'gemini',
|
||||
mode: 'write',
|
||||
},
|
||||
},
|
||||
{
|
||||
name: 'Shell-like patterns',
|
||||
input: {
|
||||
prompt: `Run: npm run build && npm test | grep "passed"
|
||||
Expected output: All tests passed`,
|
||||
tool: 'gemini',
|
||||
mode: 'analysis',
|
||||
},
|
||||
},
|
||||
];
|
||||
|
||||
/**
|
||||
* Simulate prompt processing (mirrors cli.ts logic)
|
||||
*/
|
||||
function simulatePromptProcessing(input) {
|
||||
const { prompt, tool, mode, rule = 'universal-rigorous-style' } = input;
|
||||
|
||||
// Step 1: Get base prompt
|
||||
let finalPrompt = prompt;
|
||||
|
||||
// Step 2: Extract @ patterns from CONTEXT (for cache simulation)
|
||||
const contextMatch = prompt.match(/CONTEXT:\s*([^\n]+)/i);
|
||||
let extractedPatterns = [];
|
||||
if (contextMatch) {
|
||||
const contextLine = contextMatch[1];
|
||||
const patternMatches = contextLine.matchAll(/@[^\s|]+/g);
|
||||
extractedPatterns = Array.from(patternMatches).map(m => m[0]);
|
||||
}
|
||||
|
||||
// Step 3: Simulate template concatenation
|
||||
const mockSystemRules = `[SYSTEM RULES - ${mode} mode protocol loaded]`;
|
||||
const mockRoles = `[ROLES - ${rule} template loaded]`;
|
||||
|
||||
const parts = [finalPrompt];
|
||||
parts.push(`\n=== SYSTEM RULES ===\n${mockSystemRules}`);
|
||||
parts.push(`\n=== ROLES ===\n${mockRoles}`);
|
||||
|
||||
return {
|
||||
originalPrompt: prompt,
|
||||
finalPrompt: parts.join('\n'),
|
||||
extractedPatterns,
|
||||
metadata: {
|
||||
tool,
|
||||
mode,
|
||||
rule,
|
||||
originalLength: prompt.length,
|
||||
finalLength: parts.join('\n').length,
|
||||
lineCount: parts.join('\n').split('\n').length,
|
||||
hasMultiline: prompt.includes('\n'),
|
||||
hasAtPatterns: extractedPatterns.length > 0,
|
||||
},
|
||||
};
|
||||
}
|
||||
|
||||
/**
|
||||
* Display test result
|
||||
*/
|
||||
function displayResult(testCase, result) {
|
||||
console.log(chalk.bold.cyan('\n' + '═'.repeat(70)));
|
||||
console.log(chalk.bold.white(`📋 Test: ${testCase.name}`));
|
||||
console.log(chalk.cyan('═'.repeat(70)));
|
||||
|
||||
// Input section
|
||||
console.log(chalk.bold.yellow('\n📥 INPUT:'));
|
||||
console.log(chalk.gray(' Tool: ') + chalk.green(testCase.input.tool));
|
||||
console.log(chalk.gray(' Mode: ') + chalk.green(testCase.input.mode));
|
||||
console.log(chalk.gray(' Prompt:'));
|
||||
console.log(chalk.white(' ┌' + '─'.repeat(66) + '┐'));
|
||||
testCase.input.prompt.split('\n').forEach(line => {
|
||||
const truncated = line.length > 64 ? line.substring(0, 61) + '...' : line;
|
||||
console.log(chalk.white(' │ ') + chalk.cyan(truncated.padEnd(64)) + chalk.white(' │'));
|
||||
});
|
||||
console.log(chalk.white(' └' + '─'.repeat(66) + '┘'));
|
||||
|
||||
// Metadata section
|
||||
console.log(chalk.bold.yellow('\n📊 METADATA:'));
|
||||
console.log(chalk.gray(' Original length: ') + chalk.magenta(result.metadata.originalLength + ' chars'));
|
||||
console.log(chalk.gray(' Final length: ') + chalk.magenta(result.metadata.finalLength + ' chars'));
|
||||
console.log(chalk.gray(' Line count: ') + chalk.magenta(result.metadata.lineCount));
|
||||
console.log(chalk.gray(' Has multiline: ') + (result.metadata.hasMultiline ? chalk.green('✓') : chalk.red('✗')));
|
||||
console.log(chalk.gray(' Has @ patterns: ') + (result.metadata.hasAtPatterns ? chalk.green('✓') : chalk.red('✗')));
|
||||
|
||||
if (result.extractedPatterns.length > 0) {
|
||||
console.log(chalk.gray(' Extracted patterns:'));
|
||||
result.extractedPatterns.forEach(p => {
|
||||
console.log(chalk.gray(' • ') + chalk.blue(p));
|
||||
});
|
||||
}
|
||||
|
||||
// Final prompt section
|
||||
console.log(chalk.bold.yellow('\n📤 FINAL PROMPT (passed to CLI):'));
|
||||
console.log(chalk.white(' ┌' + '─'.repeat(66) + '┐'));
|
||||
result.finalPrompt.split('\n').slice(0, 15).forEach(line => {
|
||||
const truncated = line.length > 64 ? line.substring(0, 61) + '...' : line;
|
||||
console.log(chalk.white(' │ ') + chalk.green(truncated.padEnd(64)) + chalk.white(' │'));
|
||||
});
|
||||
if (result.finalPrompt.split('\n').length > 15) {
|
||||
console.log(chalk.white(' │ ') + chalk.dim(`... (${result.finalPrompt.split('\n').length - 15} more lines)`.padEnd(64)) + chalk.white(' │'));
|
||||
}
|
||||
console.log(chalk.white(' └' + '─'.repeat(66) + '┘'));
|
||||
}
|
||||
|
||||
/**
|
||||
* Main execution
|
||||
*/
|
||||
function main() {
|
||||
console.log(chalk.bold.magenta('\n' + '█'.repeat(70)));
|
||||
console.log(chalk.bold.white(' CLI PROMPT SIMULATION - Testing Different Prompt Formats'));
|
||||
console.log(chalk.bold.magenta('█'.repeat(70)));
|
||||
|
||||
console.log(chalk.gray('\nThis script simulates how different prompt formats are processed'));
|
||||
console.log(chalk.gray('and shows the final content passed to the CLI executor.\n'));
|
||||
|
||||
for (const testCase of testCases) {
|
||||
const result = simulatePromptProcessing(testCase.input);
|
||||
displayResult(testCase, result);
|
||||
}
|
||||
|
||||
console.log(chalk.bold.magenta('\n' + '█'.repeat(70)));
|
||||
console.log(chalk.bold.white(` Completed ${testCases.length} simulations`));
|
||||
console.log(chalk.bold.magenta('█'.repeat(70) + '\n'));
|
||||
}
|
||||
|
||||
main();
|
||||
159
ccw/scripts/test-shell-prompt.js
Normal file
159
ccw/scripts/test-shell-prompt.js
Normal file
@@ -0,0 +1,159 @@
|
||||
#!/usr/bin/env node
|
||||
/**
|
||||
* CLI Shell Prompt Test Script
|
||||
*
|
||||
* Tests actual shell execution of different prompt formats.
|
||||
* Demonstrates correct vs incorrect multi-line prompt handling.
|
||||
*/
|
||||
|
||||
import { execSync, exec } from 'child_process';
|
||||
import { writeFileSync, unlinkSync, mkdtempSync } from 'fs';
|
||||
import { join } from 'path';
|
||||
import { tmpdir } from 'os';
|
||||
import chalk from 'chalk';
|
||||
|
||||
const tmpDir = mkdtempSync(join(tmpdir(), 'ccw-shell-test-'));
|
||||
|
||||
/**
|
||||
* Execute ccw cli test-parse and capture output
|
||||
*/
|
||||
function testParse(command, description) {
|
||||
console.log(chalk.bold.cyan(`\n${'─'.repeat(70)}`));
|
||||
console.log(chalk.bold.white(`📋 ${description}`));
|
||||
console.log(chalk.gray(`Command: ${command.substring(0, 100)}${command.length > 100 ? '...' : ''}`));
|
||||
console.log(chalk.cyan('─'.repeat(70)));
|
||||
|
||||
try {
|
||||
const result = execSync(command, {
|
||||
encoding: 'utf8',
|
||||
cwd: process.cwd(),
|
||||
shell: true,
|
||||
timeout: 10000,
|
||||
});
|
||||
|
||||
// Extract key info from output
|
||||
const promptMatch = result.match(/Value: "([^"]+)"/);
|
||||
const sourceMatch = result.match(/Source: ([^\n]+)/);
|
||||
|
||||
if (promptMatch) {
|
||||
console.log(chalk.green('✓ Parsed prompt: ') + chalk.yellow(promptMatch[1].substring(0, 60)));
|
||||
}
|
||||
if (sourceMatch) {
|
||||
console.log(chalk.gray(' Source: ') + sourceMatch[1]);
|
||||
}
|
||||
|
||||
// Check if prompt was truncated or split
|
||||
if (result.includes('Positional Arguments') && result.includes('[0]:')) {
|
||||
const posMatch = result.match(/\[0\]: "([^"]+)"/);
|
||||
if (posMatch) {
|
||||
console.log(chalk.red('⚠ WARNING: Part of prompt leaked to positional args: ') + chalk.yellow(posMatch[1]));
|
||||
}
|
||||
}
|
||||
|
||||
return { success: true, output: result };
|
||||
} catch (error) {
|
||||
console.log(chalk.red('✗ Error: ') + error.message);
|
||||
return { success: false, error };
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Main test suite
|
||||
*/
|
||||
function main() {
|
||||
console.log(chalk.bold.magenta('\n' + '█'.repeat(70)));
|
||||
console.log(chalk.bold.white(' SHELL PROMPT FORMAT TESTS'));
|
||||
console.log(chalk.bold.magenta('█'.repeat(70)));
|
||||
|
||||
// ============================================
|
||||
// INCORRECT METHODS (will fail or parse wrong)
|
||||
// ============================================
|
||||
console.log(chalk.bold.red('\n\n⛔ INCORRECT METHODS (will fail):'));
|
||||
|
||||
// Method 1: Direct multi-line in double quotes (WRONG)
|
||||
testParse(
|
||||
`ccw cli test-parse -p "PURPOSE: Test
|
||||
TASK: Step 1" --tool gemini`,
|
||||
'❌ Direct multi-line in double quotes (WRONG)'
|
||||
);
|
||||
|
||||
// ============================================
|
||||
// CORRECT METHODS
|
||||
// ============================================
|
||||
console.log(chalk.bold.green('\n\n✅ CORRECT METHODS:'));
|
||||
|
||||
// Method 1: Single line (works)
|
||||
testParse(
|
||||
`ccw cli test-parse -p "PURPOSE: Test | TASK: Step 1" --tool gemini`,
|
||||
'✓ Single line with pipe separator'
|
||||
);
|
||||
|
||||
// Method 2: File-based (-f option) - RECOMMENDED
|
||||
const promptFile = join(tmpDir, 'prompt.txt');
|
||||
writeFileSync(promptFile, `PURPOSE: Identify vulnerabilities
|
||||
TASK: • Scan injection flaws • Check auth bypass
|
||||
MODE: analysis
|
||||
CONTEXT: @src/auth/**/*
|
||||
EXPECTED: Security report`);
|
||||
|
||||
testParse(
|
||||
`ccw cli test-parse -f "${promptFile}" --tool gemini`,
|
||||
'✓ File-based prompt (-f option) - RECOMMENDED'
|
||||
);
|
||||
|
||||
// Method 3: $'...' syntax with literal \n (bash only)
|
||||
testParse(
|
||||
`ccw cli test-parse -p $'PURPOSE: Test\\nTASK: Step 1\\nMODE: analysis' --tool gemini`,
|
||||
"✓ $'...' syntax with \\n (bash only)"
|
||||
);
|
||||
|
||||
// Method 4: Heredoc via stdin (if supported)
|
||||
// Note: ccw cli currently doesn't support stdin, but showing the pattern
|
||||
console.log(chalk.bold.cyan(`\n${'─'.repeat(70)}`));
|
||||
console.log(chalk.bold.white(`📋 ✓ Heredoc pattern (for reference)`));
|
||||
console.log(chalk.gray(`Command: cat <<'EOF' > prompt.txt && ccw cli -f prompt.txt ...`));
|
||||
console.log(chalk.cyan('─'.repeat(70)));
|
||||
console.log(chalk.green('✓ Create file with heredoc, then use -f option'));
|
||||
|
||||
// Method 5: Single line with escaped newlines in content
|
||||
testParse(
|
||||
`ccw cli test-parse -p "PURPOSE: Test | TASK: Step 1 | MODE: analysis | CONTEXT: @src/**/*" --tool gemini`,
|
||||
'✓ Single line with | as logical separator'
|
||||
);
|
||||
|
||||
// Cleanup
|
||||
try {
|
||||
unlinkSync(promptFile);
|
||||
} catch {}
|
||||
|
||||
// Summary
|
||||
console.log(chalk.bold.magenta('\n\n' + '█'.repeat(70)));
|
||||
console.log(chalk.bold.white(' SUMMARY: Recommended Approaches'));
|
||||
console.log(chalk.bold.magenta('█'.repeat(70)));
|
||||
|
||||
console.log(`
|
||||
${chalk.bold.green('For multi-line prompts, use ONE of these methods:')}
|
||||
|
||||
${chalk.bold('1. File-based (RECOMMENDED):')}
|
||||
${chalk.cyan('ccw cli -f prompt.txt --tool gemini --mode analysis')}
|
||||
|
||||
${chalk.bold("2. $'...' syntax (bash only):")}
|
||||
${chalk.cyan("ccw cli -p $'PURPOSE: ...\\nTASK: ...\\nMODE: ...' --tool gemini")}
|
||||
|
||||
${chalk.bold('3. Heredoc + file:')}
|
||||
${chalk.cyan(`cat <<'EOF' > /tmp/prompt.txt
|
||||
PURPOSE: ...
|
||||
TASK: ...
|
||||
EOF
|
||||
ccw cli -f /tmp/prompt.txt --tool gemini`)}
|
||||
|
||||
${chalk.bold('4. Single line with separators:')}
|
||||
${chalk.cyan('ccw cli -p "PURPOSE: ... | TASK: ... | MODE: ..." --tool gemini')}
|
||||
|
||||
${chalk.bold.red('AVOID:')}
|
||||
${chalk.red('ccw cli -p "PURPOSE: ...')}
|
||||
${chalk.red('TASK: ..." # Multi-line in quotes - WILL BREAK')}
|
||||
`);
|
||||
}
|
||||
|
||||
main();
|
||||
Reference in New Issue
Block a user