@stackmemoryai/stackmemory 1.5.9 → 1.6.1

package/scripts/gepa/optimize.js

@@ -1,853 +0,0 @@
-#!/usr/bin/env node
-/**
- * GEPA Optimizer
- *
- * Genetic Eval-driven Prompt Algorithm for optimizing CLAUDE.md
- *
- * Usage:
- *   node optimize.js init                    # Initialize with current CLAUDE.md
- *   node optimize.js mutate                  # Generate new variants
- *   node optimize.js eval [variant]          # Run evals on variant(s)
- *   node optimize.js score                   # Score all variants in generation
- *   node optimize.js select                  # Select best, advance generation
- *   node optimize.js run [generations]       # Full optimization loop
- *   node optimize.js status                  # Show current status
- *   node optimize.js diff [a] [b]            # Compare two variants
- */
-
-import fs from 'fs';
-import path from 'path';
-import { fileURLToPath } from 'url';
-import { spawn, execSync } from 'child_process';
-
-const __dirname = path.dirname(fileURLToPath(import.meta.url));
-const PROJECT_ROOT = path.resolve(__dirname, '../..');
-
-// Load .env from project root
-const envPath = path.join(PROJECT_ROOT, '.env');
-if (fs.existsSync(envPath)) {
-  for (const line of fs.readFileSync(envPath, 'utf8').split('\n')) {
-    const match = line.match(/^([A-Z_][A-Z0-9_]*)=(.*)$/);
-    if (match && !process.env[match[1]]) {
-      process.env[match[1]] = match[2].replace(/^["']|["']$/g, '');
-    }
-  }
-}
-
-// Configuration
-const CONFIG_PATH = path.join(__dirname, 'config.json');
-const config = JSON.parse(fs.readFileSync(CONFIG_PATH, 'utf8'));
-
-const GEPA_DIR = process.env.GEPA_DIR || __dirname;
-const GENERATIONS_DIR = path.join(GEPA_DIR, 'generations');
-const RESULTS_DIR = path.join(GEPA_DIR, 'results');
-const EVALS_DIR = path.join(GEPA_DIR, 'evals');
-
-// Ensure directories
-[GENERATIONS_DIR, RESULTS_DIR, EVALS_DIR].forEach((dir) => {
-  if (!fs.existsSync(dir)) fs.mkdirSync(dir, { recursive: true });
-});
-
-/**
- * State management
- */
-function getState() {
-  const statePath = path.join(GEPA_DIR, 'state.json');
-  if (fs.existsSync(statePath)) {
-    return JSON.parse(fs.readFileSync(statePath, 'utf8'));
-  }
-  return {
-    currentGeneration: 0,
-    bestVariant: null,
-    bestScore: 0,
-    history: [],
-  };
-}
-
-function saveState(state) {
-  fs.writeFileSync(
-    path.join(GEPA_DIR, 'state.json'),
-    JSON.stringify(state, null, 2)
-  );
-}
-
-/**
- * Get path for a generation/variant
- */
-function getGenPath(gen, variant = null) {
-  const genDir = path.join(
-    GENERATIONS_DIR,
-    `gen-${String(gen).padStart(3, '0')}`
-  );
-  if (!fs.existsSync(genDir)) fs.mkdirSync(genDir, { recursive: true });
-  return variant ? path.join(genDir, `${variant}.md`) : genDir;
-}
-
-/**
- * Initialize GEPA with current CLAUDE.md
- */
-async function init(targetPath) {
-  const claudeMdPath = targetPath || path.join(process.cwd(), 'CLAUDE.md');
-
-  if (!fs.existsSync(claudeMdPath)) {
-    console.error(`Error: ${claudeMdPath} not found`);
-    process.exit(1);
-  }
-
-  const content = fs.readFileSync(claudeMdPath, 'utf8');
-  const genPath = getGenPath(0, 'baseline');
-
-  fs.writeFileSync(genPath, content);
-
-  const state = {
-    currentGeneration: 0,
-    bestVariant: 'baseline',
-    bestScore: 0,
-    targetPath: claudeMdPath,
-    history: [
-      {
-        generation: 0,
-        variant: 'baseline',
-        action: 'init',
-        timestamp: new Date().toISOString(),
-      },
-    ],
-  };
-
-  saveState(state);
-  console.log(`Initialized GEPA with ${claudeMdPath}`);
-  console.log(`Baseline saved to ${genPath}`);
-}
-
-/**
- * Generate mutations of the current best variant
- */
-async function mutate() {
-  const state = getState();
-  const nextGen = state.currentGeneration + 1;
-  const currentBest = fs.readFileSync(
-    getGenPath(state.currentGeneration, state.bestVariant),
-    'utf8'
-  );
-
-  console.log(
-    `Generating ${config.evolution.populationSize} variants for generation ${nextGen}...`
-  );
-
-  const mutations = config.evolution.mutationStrategies;
-  const variants = [];
-
-  for (let i = 0; i < config.evolution.populationSize; i++) {
-    const strategy = mutations[i % mutations.length];
-    const variantName = `variant-${String.fromCharCode(97 + i)}`; // a, b, c, d...
-
-    console.log(`  Creating ${variantName} using strategy: ${strategy}`);
-
-    const mutatedContent = await generateMutation(currentBest, strategy, state);
-    const variantPath = getGenPath(nextGen, variantName);
-
-    fs.writeFileSync(variantPath, mutatedContent);
-    variants.push({ name: variantName, strategy, path: variantPath });
-  }
-
-  // Also copy baseline for comparison
-  fs.writeFileSync(getGenPath(nextGen, 'baseline'), currentBest);
-
-  state.history.push({
-    generation: nextGen,
-    action: 'mutate',
-    variants: variants.map((v) => v.name),
-    timestamp: new Date().toISOString(),
-  });
-  saveState(state);
-
-  console.log(
-    `\nGenerated ${variants.length} variants in gen-${String(nextGen).padStart(3, '0')}/`
-  );
-  return variants;
-}
-
-/**
- * Generate a mutation using AI
- */
-async function generateMutation(content, strategy, state) {
-  const strategyPrompts = {
-    rephrase: `Rephrase instructions for clarity without changing meaning. Make them more direct and actionable.`,
-
-    add_examples: `Add concrete examples where instructions are abstract. Use <example> tags for code examples.`,
-
-    remove_redundancy: `Remove redundant or repetitive instructions. Consolidate similar rules. Keep it DRY.`,
-
-    restructure: `Reorganize sections for better flow. Group related instructions. Improve hierarchy.`,
-
-    add_constraints: `Add specific constraints and guardrails based on common failure modes. Be precise about what NOT to do.`,
-
-    simplify: `Simplify complex instructions. Break down multi-step rules. Use bullet points over paragraphs.`,
-  };
-
-  const prompt = `You are optimizing a CLAUDE.md system prompt for an AI coding agent.
-
-CURRENT PROMPT:
-\`\`\`markdown
-${content}
-\`\`\`
-
-OPTIMIZATION STRATEGY: ${strategy}
-${strategyPrompts[strategy]}
-
-EVALUATION FEEDBACK FROM PREVIOUS GENERATIONS:
-${getRecentFeedback(state)}
-
-REFLECTION INSIGHTS (from failure pattern analysis):
-${getReflectionInsights()}
-
-REQUIREMENTS:
-1. Output ONLY the improved markdown content
-2. Preserve all critical instructions and constraints
-3. Keep the same overall structure unless restructuring
-4. Do not add commentary or explanations
-5. Target <8000 tokens total length
-
-OUTPUT THE IMPROVED CLAUDE.MD:`;
-
-  // Use Claude to generate mutation
-  const result = await callClaude(prompt);
-  return result.trim();
-}
-
-/**
- * Get recent evaluation feedback for context
- */
-function getRecentFeedback(state) {
-  const scoresPath = path.join(RESULTS_DIR, 'scores.jsonl');
-  if (!fs.existsSync(scoresPath)) return 'No previous evaluations.';
-
-  const lines = fs
-    .readFileSync(scoresPath, 'utf8')
-    .trim()
-    .split('\n')
-    .slice(-20);
-  const scores = lines.map((l) => JSON.parse(l));
-
-  const summary = scores.reduce((acc, s) => {
-    if (!acc[s.variant]) acc[s.variant] = { total: 0, count: 0, errors: 0 };
-    acc[s.variant].total += s.metrics?.successfulToolCalls || 0;
-    acc[s.variant].count++;
-    acc[s.variant].errors += s.metrics?.errorCount || 0;
-    return acc;
-  }, {});
-
-  return Object.entries(summary)
-    .map(
-      ([v, s]) =>
-        `${v}: ${s.count} sessions, ${s.errors} errors, avg success: ${(s.total / s.count).toFixed(1)}`
-    )
-    .join('\n');
-}
-
-/**
- * Load most recent reflection insights for mutation context
- */
-function getReflectionInsights() {
-  const reflectionFiles = fs.existsSync(RESULTS_DIR)
-    ? fs
-        .readdirSync(RESULTS_DIR)
-        .filter((f) => f.startsWith('reflection-') && f.endsWith('.json'))
-    : [];
-
-  if (reflectionFiles.length === 0) return 'No reflection data yet.';
-
-  // Pick the most recent reflection file
-  reflectionFiles.sort().reverse();
-  const latest = JSON.parse(
-    fs.readFileSync(path.join(RESULTS_DIR, reflectionFiles[0]), 'utf8')
-  );
-
-  const insights = latest.insights;
-  if (!insights) return 'No reflection insights available.';
-
-  const parts = [];
-
-  if (insights.failureModes?.length) {
-    parts.push(`Failure modes: ${insights.failureModes.join('; ')}`);
-  }
-  if (insights.missingInstructions?.length) {
-    parts.push(
-      `Missing instructions: ${insights.missingInstructions.join('; ')}`
-    );
-  }
-  if (insights.unclearInstructions?.length) {
-    parts.push(
-      `Unclear instructions: ${insights.unclearInstructions.join('; ')}`
-    );
-  }
-  if (insights.priorityMutations?.length) {
-    parts.push(
-      `Priority changes:\n${insights.priorityMutations
-        .map((m) => `  - [${m.type}] ${m.section}: ${m.change}`)
-        .join('\n')}`
-    );
-  }
-
-  return parts.join('\n') || 'No actionable insights.';
-}
-
-/**
- * Call Claude CLI via spawn (stdin pipe, no shell interpolation)
- */
-function spawnClaude(prompt, { cwd, timeoutMs } = {}) {
-  return new Promise((resolve, reject) => {
-    const args = ['--print'];
-    const child = spawn('claude', args, {
-      cwd,
-      stdio: ['pipe', 'pipe', 'pipe'],
-      env: { ...process.env },
-    });
-
-    let stdout = '';
-    let stderr = '';
-    let killed = false;
-
-    const timer = timeoutMs
-      ? setTimeout(() => {
-          killed = true;
-          child.kill('SIGTERM');
-        }, timeoutMs)
-      : null;
-
-    child.stdout.on('data', (d) => (stdout += d));
-    child.stderr.on('data', (d) => (stderr += d));
-
-    child.on('close', (code) => {
-      if (timer) clearTimeout(timer);
-      if (killed)
-        return reject(new Error(`claude timed out after ${timeoutMs}ms`));
-      if (code !== 0 && !stdout)
-        return reject(new Error(stderr || `claude exited ${code}`));
-      resolve(stdout);
-    });
-
-    child.on('error', (err) => {
-      if (timer) clearTimeout(timer);
-      reject(err);
-    });
-
-    child.stdin.write(prompt);
-    child.stdin.end();
-  });
-}
-
-/**
- * Call Claude API for mutation generation
- */
-async function callClaude(prompt) {
-  // Try using claude CLI first (stdin pipe, no shell injection)
-  try {
-    return await spawnClaude(prompt);
-  } catch (e) {
-    // Fallback to API
-    const apiKey = process.env.ANTHROPIC_API_KEY;
-    if (!apiKey) {
-      console.error('Error: ANTHROPIC_API_KEY not set and claude CLI failed');
-      process.exit(1);
-    }
-
-    const response = await fetch('https://api.anthropic.com/v1/messages', {
-      method: 'POST',
-      headers: {
-        'Content-Type': 'application/json',
-        'x-api-key': apiKey,
-        'anthropic-version': '2023-06-01',
-      },
-      body: JSON.stringify({
-        model: 'claude-sonnet-4-20250514',
-        max_tokens: 8000,
-        messages: [{ role: 'user', content: prompt }],
-      }),
-    });
-
-    const data = await response.json();
-    return data.content[0].text;
-  }
-}
-
-/**
- * Run evaluations on a variant
- */
-async function runEval(variantName) {
-  const state = getState();
-  const gen = state.currentGeneration + 1; // Eval next gen variants
-  const variantPath = getGenPath(gen, variantName);
-
-  if (!fs.existsSync(variantPath)) {
-    console.error(`Variant not found: ${variantPath}`);
-    return null;
-  }
-
-  console.log(`Running evals on ${variantName}...`);
-
-  // Load eval tasks
-  const evalFiles = fs
-    .readdirSync(EVALS_DIR)
-    .filter((f) => f.endsWith('.jsonl'));
-  const tasks = evalFiles.flatMap((f) =>
-    fs
-      .readFileSync(path.join(EVALS_DIR, f), 'utf8')
-      .trim()
-      .split('\n')
-      .map((l) => JSON.parse(l))
-  );
-
-  console.log(`  Found ${tasks.length} eval tasks`);
-
-  // Set environment for tracking
-  process.env.GEPA_VARIANT = variantName;
-  process.env.GEPA_GENERATION = String(gen);
-
-  const results = [];
-
-  for (const task of tasks.slice(0, config.evals.minSamplesPerVariant)) {
-    console.log(`  Running: ${task.name}`);
-
-    const result = await runSingleEval(task, variantPath);
-    results.push({
-      taskId: task.id,
-      taskName: task.name,
-      weight: task.weight || 1.0,
-      ...result,
-    });
-  }
-
-  // Save results
-  const resultsPath = path.join(RESULTS_DIR, `eval-${gen}-${variantName}.json`);
-  fs.writeFileSync(
-    resultsPath,
-    JSON.stringify({ variant: variantName, generation: gen, results }, null, 2)
-  );
-
-  // Calculate aggregate score
-  const score = calculateScore(results);
-  console.log(`  Score: ${(score * 100).toFixed(1)}%`);
-
-  return { variant: variantName, score, results };
-}
-
-/**
- * Run a single eval task
- */
-async function runSingleEval(task, variantPath) {
-  const startTime = Date.now();
-  let tempDir;
-
-  try {
-    // Create temp project with variant as CLAUDE.md
-    tempDir = fs.mkdtempSync('/tmp/gepa-eval-');
-    fs.copyFileSync(variantPath, path.join(tempDir, 'CLAUDE.md'));
-
-    // Copy fixture if needed
-    if (task.input_file) {
-      const fixturePath = path.join(EVALS_DIR, task.input_file);
-      if (fs.existsSync(fixturePath)) {
-        fs.copyFileSync(
-          fixturePath,
-          path.join(tempDir, path.basename(task.input_file))
-        );
-      }
-    }
-
-    // Run claude via spawn with Node-native timeout (no GNU timeout needed)
-    const result = await spawnClaude(task.prompt, {
-      cwd: tempDir,
-      timeoutMs: config.evals.timeout,
-    });
-
-    // Evaluate result against expected outcomes (LLM judge with regex fallback)
-    const evaluation = await evaluateExpectations(result, task.expected, task);
-
-    return {
-      passed: evaluation.passed,
-      passRate: evaluation.passRate,
-      criteria: evaluation.criteria,
-      judgeMode: evaluation.judgeMode,
-      duration: Date.now() - startTime,
-      output: result.slice(0, 2000),
-    };
-  } catch (error) {
-    return {
-      passed: false,
-      duration: Date.now() - startTime,
-      error: error.message,
-    };
-  } finally {
-    if (tempDir) fs.rmSync(tempDir, { recursive: true, force: true });
-  }
-}
-
-/**
- * Evaluate output against expectations using LLM judge (regex fallback)
- */
-async function evaluateExpectations(output, expected, task) {
-  if (!expected)
-    return { passed: true, passRate: 1.0, criteria: {}, judgeMode: 'skip' };
-
-  // Try LLM judge first
-  try {
-    const result = await llmJudge(output, expected, task);
-    return { ...result, judgeMode: 'llm' };
-  } catch (e) {
-    console.log(`    LLM judge failed (${e.message}), using regex fallback`);
-    const result = regexJudge(output, expected);
-    return { ...result, judgeMode: 'regex' };
-  }
-}
-
-/**
- * LLM-as-judge — uses a fast model to evaluate output against criteria
- */
-async function llmJudge(output, expected, task) {
-  const criteriaList = Object.entries(expected)
-    .map(
-      ([key, value]) =>
-        `- ${key}: ${typeof value === 'string' ? value : 'should be ' + value}`
-    )
-    .join('\n');
-
-  const judgePrompt = `You are a strict code evaluation judge. Evaluate whether the AI output satisfies each criterion.
-
-TASK GIVEN TO AI:
-${task.prompt}
-
-AI OUTPUT:
-\`\`\`
-${output.slice(0, 6000)}
-\`\`\`
-
-CRITERIA TO EVALUATE:
-${criteriaList}
-
-For each criterion, determine if the output genuinely satisfies it. Be strict:
-- "has_function" means a real, working function is defined (not just mentioned)
-- "bug_fixed" means the actual bug is corrected (not just discussed)
-- "handles_edge_cases" means edge cases are actually handled in code
-- "explains_fix" means there's a clear explanation of what was wrong and why
-
-Respond with ONLY this JSON (no markdown fences):
-{
-  "criteria": {
-    "criterion_name": {"passed": true, "reason": "brief explanation"},
-    "criterion_name": {"passed": false, "reason": "brief explanation"}
-  }
-}`;
-
-  const judgeModel = config.judge?.model || 'claude-haiku-4-5-20251001';
-  const raw = await callJudge(judgePrompt, judgeModel);
-
-  // Extract JSON from response
-  const jsonMatch = raw.match(/\{[\s\S]*\}/);
-  if (!jsonMatch) throw new Error('No JSON in judge response');
-
-  const parsed = JSON.parse(jsonMatch[0]);
-  const criteria = parsed.criteria || {};
-  const entries = Object.values(criteria);
-  const passedCount = entries.filter((c) => c.passed).length;
-  const passRate = entries.length > 0 ? passedCount / entries.length : 0;
-
-  return {
-    passed: passRate >= 0.6,
-    passRate,
-    criteria,
-  };
-}
-
-/**
- * Call judge model via Anthropic API (fast, cheap model for evaluation)
- */
-async function callJudge(prompt, model) {
-  const apiKey = process.env.ANTHROPIC_API_KEY;
-
-  if (apiKey) {
-    const response = await fetch('https://api.anthropic.com/v1/messages', {
-      method: 'POST',
-      headers: {
-        'Content-Type': 'application/json',
-        'x-api-key': apiKey,
-        'anthropic-version': '2023-06-01',
-      },
-      body: JSON.stringify({
-        model,
-        max_tokens: 2000,
-        messages: [{ role: 'user', content: prompt }],
-      }),
-    });
-
-    if (!response.ok) {
-      throw new Error(`Judge API error: ${response.status}`);
-    }
-
-    const data = await response.json();
-    return data.content[0].text;
-  }
-
-  // Fallback to CLI
-  return await spawnClaude(prompt, { timeoutMs: 30000 });
-}
-
-/**
- * Regex fallback judge (used when LLM judge is unavailable)
- */
-function regexJudge(output, expected) {
-  const criteria = {};
-
-  for (const [key, value] of Object.entries(expected)) {
-    let passed = false;
-    switch (key) {
-      case 'has_function':
-        passed =
-          /function\s+\w+|const\s+\w+\s*=\s*(\([^)]*\)|async)?\s*(=>|\{)/.test(
-            output
-          );
-        break;
-      case 'handles_edge_cases':
-        passed = /if\s*\(|edge|empty|null|undefined|\.length/.test(output);
-        break;
-      case 'uses_async':
-        passed = /async|await|Promise/.test(output);
-        break;
-      case 'no_nested_callbacks':
-        passed = !/callback\s*\(\s*function|\.then\s*\([^)]*\.then/.test(
-          output
-        );
-        break;
-      case 'bug_fixed':
-        passed = /fix|correct|change|update/i.test(output);
-        break;
-      case 'explains_fix':
-        passed =
-          output.length > 200 &&
-          /because|since|the issue|the problem/i.test(output);
-        break;
-      default:
-        passed = output.toLowerCase().includes(key.toLowerCase());
-    }
-    criteria[key] = { passed, reason: 'regex heuristic' };
-  }
-
-  const entries = Object.values(criteria);
-  const passedCount = entries.filter((c) => c.passed).length;
-  const passRate = entries.length > 0 ? passedCount / entries.length : 0;
-
-  return { passed: passRate >= 0.6, passRate, criteria };
-}
-
-/**
- * Calculate weighted score using task weights and per-criterion pass rates
- */
-function calculateScore(results) {
-  if (results.length === 0) return 0;
-
-  let weightedSum = 0;
-  let totalWeight = 0;
-
-  for (const r of results) {
-    const weight = r.weight || 1.0;
-    // Use passRate for partial credit when available (LLM judge),
-    // fall back to binary passed for regex judge
-    const score = r.passRate !== undefined ? r.passRate : r.passed ? 1.0 : 0.0;
-    weightedSum += score * weight;
-    totalWeight += weight;
-  }
-
-  return totalWeight > 0 ? weightedSum / totalWeight : 0;
-}
-
-/**
- * Score all variants and select best
- */
-async function scoreAndSelect() {
-  const state = getState();
-  const gen = state.currentGeneration + 1;
-  const genDir = getGenPath(gen);
-
-  if (!fs.existsSync(genDir)) {
-    console.error(`Generation ${gen} not found. Run 'mutate' first.`);
-    return;
-  }
-
-  const variants = fs
-    .readdirSync(genDir)
-    .filter((f) => f.endsWith('.md'))
-    .map((f) => f.replace('.md', ''));
-
-  console.log(`Scoring ${variants.length} variants in generation ${gen}...`);
-
-  const scores = [];
-
-  for (const variant of variants) {
-    const result = await runEval(variant);
-    if (result) scores.push(result);
-  }
-
-  // Sort by score
-  scores.sort((a, b) => b.score - a.score);
-
-  console.log('\nResults:');
-  scores.forEach((s, i) => {
-    const marker = i === 0 ? ' <-- BEST' : '';
-    console.log(
-      `  ${i + 1}. ${s.variant}: ${(s.score * 100).toFixed(1)}%${marker}`
-    );
-  });
-
-  // Select best
-  const best = scores[0];
-
-  if (best.score > state.bestScore) {
-    state.currentGeneration = gen;
-    state.bestVariant = best.variant;
-    state.bestScore = best.score;
-
-    // Update symlink
-    const currentLink = path.join(GENERATIONS_DIR, 'current');
-    if (fs.existsSync(currentLink)) fs.unlinkSync(currentLink);
-    fs.symlinkSync(getGenPath(gen, best.variant), currentLink);
-
-    console.log(
-      `\nNew best: ${best.variant} (${(best.score * 100).toFixed(1)}%)`
-    );
-    console.log(
-      `Symlink updated: generations/current -> gen-${String(gen).padStart(3, '0')}/${best.variant}.md`
-    );
-  } else {
-    console.log(
-      `\nNo improvement over previous best (${(state.bestScore * 100).toFixed(1)}%)`
-    );
-  }
-
-  state.history.push({
-    generation: gen,
-    action: 'select',
-    scores: scores.map((s) => ({ variant: s.variant, score: s.score })),
-    best: best.variant,
-    timestamp: new Date().toISOString(),
-  });
-
-  saveState(state);
-  return best;
-}
-
-/**
- * Full optimization loop
- */
-async function run(generations = config.evolution.generations) {
-  console.log(`Starting GEPA optimization for ${generations} generations...\n`);
-
-  for (let i = 0; i < generations; i++) {
-    console.log(`\n${'='.repeat(60)}`);
-    console.log(`GENERATION ${i + 1}/${generations}`);
-    console.log(`${'='.repeat(60)}\n`);
-
-    await mutate();
-    const best = await scoreAndSelect();
-
-    if (best.score >= config.scoring.threshold) {
-      console.log(
-        `\nThreshold reached (${(config.scoring.threshold * 100).toFixed(0)}%)! Stopping early.`
-      );
-      break;
-    }
-  }
-
-  console.log('\n' + '='.repeat(60));
-  console.log('OPTIMIZATION COMPLETE');
-  console.log('='.repeat(60));
-
-  const state = getState();
-  console.log(`Best variant: ${state.bestVariant}`);
-  console.log(`Best score: ${(state.bestScore * 100).toFixed(1)}%`);
-  console.log(`Generations: ${state.currentGeneration}`);
-  console.log(`\nTo apply: cp generations/current /path/to/your/CLAUDE.md`);
-}
-
-/**
- * Show status
- */
-function status() {
-  const state = getState();
-
-  console.log('GEPA Status');
-  console.log('===========');
-  console.log(`Current generation: ${state.currentGeneration}`);
-  console.log(`Best variant: ${state.bestVariant}`);
-  console.log(`Best score: ${(state.bestScore * 100).toFixed(1)}%`);
-  console.log(`Target: ${state.targetPath}`);
-  console.log(`\nHistory:`);
-
-  state.history.slice(-10).forEach((h) => {
-    console.log(`  [${h.timestamp}] ${h.action} (gen ${h.generation})`);
-  });
-}
-
-/**
- * Diff two variants
- */
-function diff(a, b) {
-  const state = getState();
-  const gen = state.currentGeneration;
-
-  const pathA = getGenPath(gen, a || 'baseline');
-  const pathB = getGenPath(gen, b || state.bestVariant);
-
-  if (!fs.existsSync(pathA) || !fs.existsSync(pathB)) {
-    console.error('Variant not found');
-    return;
-  }
-
-  try {
-    execSync(`diff -u ${pathA} ${pathB}`, { stdio: 'inherit' });
-  } catch (e) {
-    // diff returns non-zero when files differ
-  }
-}
-
-// CLI
-const command = process.argv[2];
-const arg1 = process.argv[3];
-const arg2 = process.argv[4];
-
-switch (command) {
-  case 'init':
-    init(arg1);
-    break;
-  case 'mutate':
-    mutate();
-    break;
-  case 'eval':
-    runEval(arg1 || 'baseline');
-    break;
-  case 'select':
-  case 'score':
-    scoreAndSelect();
-    break;
-  case 'run':
-    run(parseInt(arg1) || config.evolution.generations);
-    break;
-  case 'status':
-    status();
-    break;
-  case 'diff':
-    diff(arg1, arg2);
-    break;
-  default:
-    console.log(`
-GEPA - Genetic Eval-driven Prompt Algorithm
-
-Usage:
-  node optimize.js init [path]           Initialize with CLAUDE.md
-  node optimize.js mutate                Generate new variants
-  node optimize.js eval [variant]        Run evals on variant
-  node optimize.js score                 Score all variants, select best
-  node optimize.js run [generations]     Full optimization loop
-  node optimize.js status                Show current status
-  node optimize.js diff [a] [b]          Compare two variants
-`);
-}