feat: lockfile to prevent parallel runs + AI keywords lib

job_applier.mjs

@@ -11,6 +11,7 @@ import { fileURLToPath } from 'url';
 const __dir = dirname(fileURLToPath(import.meta.url));
 
 import { getJobsByStatus, updateJobStatus, appendLog, loadConfig } from './lib/queue.mjs';
+import { acquireLock } from './lib/lock.mjs';
 import { createBrowser } from './lib/browser.mjs';
 import { FormFiller } from './lib/form_filler.mjs';
 import { verifyLogin as liLogin, applyLinkedIn } from './lib/linkedin.mjs';
@@ -25,6 +26,7 @@ import {
 const isPreview = process.argv.includes('--preview');
 
 async function main() {
+  acquireLock('applier', resolve(__dir, 'data'));
   console.log('🚀 claw-apply: Job Applier starting\n');
 
   const settings = loadConfig(resolve(__dir, 'config/settings.json'));

job_searcher.mjs

@@ -10,6 +10,7 @@ import { fileURLToPath } from 'url';
 const __dir = dirname(fileURLToPath(import.meta.url));
 
 import { addJobs, loadQueue, loadConfig } from './lib/queue.mjs';
+import { acquireLock } from './lib/lock.mjs';
 import { createBrowser } from './lib/browser.mjs';
 import { verifyLogin as liLogin, searchLinkedIn } from './lib/linkedin.mjs';
 import { verifyLogin as wfLogin, searchWellfound } from './lib/wellfound.mjs';
@@ -18,6 +19,7 @@ import { DEFAULT_FIRST_RUN_DAYS } from './lib/constants.mjs';
 import { generateKeywords } from './lib/keywords.mjs';
 
 async function main() {
+  acquireLock('searcher', resolve(__dir, 'data'));
   console.log('🔍 claw-apply: Job Searcher starting\n');
 
   // Load config

lib/keywords.mjs

@@ -0,0 +1,61 @@
+/**
+ * keywords.mjs — AI-generated search keywords
+ * One Claude call per search track using full profile + search config context
+ */
+
+export async function generateKeywords(search, profile, apiKey) {
+  if (!apiKey) throw new Error('ANTHROPIC_API_KEY not set');
+
+  const prompt = `You are an expert job search strategist helping a candidate find the right roles on LinkedIn and Wellfound.
+
+## Candidate Profile
+- Name: ${profile.name.first} ${profile.name.last}
+- Location: ${profile.location.city}, ${profile.location.state} (remote only)
+- Years experience: ${profile.years_experience}
+- Desired salary: $${profile.desired_salary.toLocaleString()}
+- Work authorization: Authorized to work in US + UK, no sponsorship needed
+- Willing to relocate: ${profile.willing_to_relocate ? 'Yes' : 'No'}
+- Background summary: ${profile.cover_letter?.substring(0, 400)}
+
+## Job Search Track: "${search.name}"
+- Salary minimum: $${(search.salary_min || 0).toLocaleString()}
+- Platforms: ${(search.platforms || []).join(', ')}
+- Remote only: ${search.filters?.remote ? 'Yes' : 'No'}
+- Exclude these keywords: ${(search.exclude_keywords || []).join(', ')}
+- Current keywords already in use: ${(search.keywords || []).join(', ')}
+
+## Task
+Generate 15 additional LinkedIn/Wellfound job search query strings to find "${search.name}" roles for this candidate.
+
+Think about:
+- How do startups and hiring managers actually title these roles at seed/Series A/B companies?
+- What variations exist across industries (fintech, devtools, data infra, security, AI/ML)?
+- What seniority + function combinations surface the best matches?
+- What terms does this specific candidate's background match well?
+- Do NOT repeat keywords already listed above
+- Do NOT use excluded keywords
+
+Return ONLY a JSON array of strings, no explanation, no markdown.
+Example format: ["query one", "query two", "query three"]`;
+
+  const res = 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-3-haiku-20240307',
+      max_tokens: 1024,
+      messages: [{ role: 'user', content: prompt }]
+    })
+  });
+
+  const data = await res.json();
+  if (data.error) throw new Error(data.error.message);
+
+  const text = data.content[0].text.trim();
+  const clean = text.replace(/```json\n?|\n?```/g, '').trim();
+  return JSON.parse(clean);
+}

lib/lock.mjs

@@ -0,0 +1,34 @@
+/**
+ * lock.mjs — Simple PID-based lockfile to prevent parallel runs
+ */
+import { existsSync, readFileSync, writeFileSync, unlinkSync } from 'fs';
+import { resolve } from 'path';
+
+export function acquireLock(name, dataDir) {
+  const lockFile = resolve(dataDir, `${name}.lock`);
+
+  if (existsSync(lockFile)) {
+    const pid = parseInt(readFileSync(lockFile, 'utf8').trim(), 10);
+    // Check if that process is still alive
+    try {
+      process.kill(pid, 0); // signal 0 = check existence only
+      console.error(`⚠️  ${name} already running (PID ${pid}). Exiting.`);
+      process.exit(0);
+    } catch {
+      // Stale lock — process is gone, safe to continue
+      console.warn(`🔓 Stale lock found (PID ${pid}), clearing.`);
+    }
+  }
+
+  writeFileSync(lockFile, String(process.pid));
+
+  // Release lock on exit (normal or crash)
+  const release = () => {
+    try { unlinkSync(lockFile); } catch {}
+  };
+  process.on('exit', release);
+  process.on('SIGINT', () => { release(); process.exit(130); });
+  process.on('SIGTERM', () => { release(); process.exit(143); });
+
+  return release;
+}

test_ai_keywords.mjs

@@ -0,0 +1,31 @@
+/**
+ * test_ai_keywords.mjs — Test AI-generated search terms with full context
+ */
+import { readFileSync } from 'fs';
+import { dirname, resolve } from 'path';
+import { fileURLToPath } from 'url';
+import { generateKeywords } from './lib/keywords.mjs';
+
+const __dir = dirname(fileURLToPath(import.meta.url));
+const profile = JSON.parse(readFileSync(resolve(__dir, 'config/profile.json'), 'utf8'));
+const searchConfig = JSON.parse(readFileSync(resolve(__dir, 'config/search_config.json'), 'utf8'));
+const apiKey = process.env.ANTHROPIC_API_KEY;
+
+async function main() {
+  console.log('🤖 AI keyword generation — full context test\n');
+
+  for (const search of searchConfig.searches) {
+    console.log(`\n━━ ${search.name} ━━`);
+    console.log('Static keywords:', search.keywords.join(', '));
+    console.log('\nGenerating AI keywords...');
+
+    const keywords = await generateKeywords(search, profile, apiKey);
+    console.log(`\nAI generated (${keywords.length}):`);
+    keywords.forEach((k, i) => console.log(`  ${i+1}. "${k}"`));
+
+    const merged = [...new Set([...search.keywords, ...keywords])];
+    console.log(`\nMerged total: ${merged.length} unique queries`);
+  }
+}
+
+main().catch(e => { console.error('Error:', e.message); process.exit(1); });