job-forge 2.0.0

package/modes/scan.md

@@ -0,0 +1,185 @@
+# Mode: scan — Portal Scanner (Offer Discovery)
+
+Scans configured job portals, filters by title relevance, and adds new offers to the pipeline for later evaluation.
+
+## Recommended Execution
+
+Run as a subagent to avoid consuming main context:
+
+```
+Agent(
+    subagent_type="general-purpose",
+    prompt="[contents of this file + specific data]",
+    run_in_background=True
+)
+```
+
+## Read This Configuration
+
+Read `portals.yml` which contains:
+- `search_queries`: List of WebSearch queries with `site:` filters per portal (broad discovery)
+- `tracked_companies`: Specific companies with `careers_url` for direct navigation
+- `title_filter`: Positive/negative/seniority_boost keywords for title filtering
+
+## Apply This Discovery Strategy (3 levels)
+
+### Use Level 1 — Direct Geometra (PRIMARY)
+
+**For each company in `tracked_companies`:** Connect to its `careers_url` with Geometra MCP (`geometra_connect` + `geometra_page_model` / `geometra_list_items`), read ALL visible job listings, and extract the title + URL of each one. Direct Geometra is the most reliable method because:
+
+- It sees the page in real time (not cached Google results).
+- It works with SPAs (Ashby, Lever, Workday).
+- It detects new offers instantly.
+- It doesn't depend on Google indexing.
+
+**Every company MUST have a `careers_url` in portals.yml.** If it doesn't, search for it once, save it, and use it in future scans.
+
+### Use Level 2 — Greenhouse API (COMPLEMENTARY)
+
+For companies using Greenhouse, the JSON API (`boards-api.greenhouse.io/v1/boards/{slug}/jobs`) returns clean structured data. Use as a quick complement to Level 1 — it's faster than Geometra but only works with Greenhouse.
+
+### Use Level 3 — WebSearch Queries (BROAD DISCOVERY)
+
+The `search_queries` with `site:` filters cover portals broadly (all Ashby boards, all Greenhouse boards, all Lever boards, all Workday boards). Useful for discovering NEW companies not yet in `tracked_companies`, but results may be outdated.
+
+**Execution priority:**
+1. Level 1: Geometra → all `tracked_companies` with `careers_url`
+2. Level 2: API → all `tracked_companies` with `api:`
+3. Level 3: WebSearch → all `search_queries` with `enabled: true`
+
+The levels are additive — all are executed, results are merged and deduplicated.
+
+## Run This Workflow
+
+1. **Read configuration**: `portals.yml`
+2. **Read history**: `data/scan-history.tsv` → previously seen URLs
+3. **Read dedup sources**: all day files in `data/applications/` + `data/pipeline.md`
+
+4. **Level 1 — Geometra scan** (sequential, or ≤2 parallel via `task` subagents per Hard Limit #1 in `AGENTS.md`):
+   For each company in `tracked_companies` with `enabled: true` and `careers_url` defined:
+   a. `geometra_connect` to the `careers_url`
+   b. `geometra_page_model` or `geometra_list_items` to read all job listings
+   c. If the page has filters/departments, navigate the relevant sections
+   d. For each job listing extract: `{title, url, company}`
+   e. If the page paginates results, navigate additional pages
+   f. Accumulate in candidates list
+   g. If `careers_url` fails (404, redirect), try `scan_query` as fallback and note for URL update
+
+5. **Level 2 — Greenhouse APIs** (WebFetch can batch freely — it's cheap and doesn't use Geometra sessions):
+   For each company in `tracked_companies` with `api:` defined and `enabled: true`:
+   a. WebFetch the API URL → JSON with job list
+   b. For each job extract: `{title, url, company}`
+   c. Accumulate in candidates list (dedup with Level 1)
+
+6. **Level 3 — WebSearch queries** (WebSearch is parallel-safe; batch freely):
+   For each query in `search_queries` with `enabled: true`:
+   a. Execute WebSearch with the defined `query`
+   b. From each result extract: `{title, url, company}`
+      - **title**: from the result title (before " @ " or " | ")
+      - **url**: result URL
+      - **company**: after " @ " in the title, or extract from domain/path
+   c. Accumulate in candidates list (dedup with Level 1+2)
+
+6. **Filter by title** using `title_filter` from `portals.yml`:
+   - At least 1 keyword from `positive` must appear in the title (case-insensitive)
+   - 0 keywords from `negative` must appear
+   - `seniority_boost` keywords give priority but are not required
+
+7. **Deduplicate** against 3 sources (URL-exact + fuzzy company+role):
+
+   **Layer 1 — URL-exact:**
+   - `scan-history.tsv` → exact URL already seen
+   - `pipeline.md` → exact URL already in pending or processed
+
+   **Layer 2 — Company + role fuzzy match (catches reposts with new URLs):**
+   - all day files in `data/applications/` → normalize company name (lowercase, strip non-alphanumeric) + fuzzy role match (2+ significant words in common, words > 3 chars). This is the same logic used in `dedup-tracker.mjs` and `merge-tracker.mjs`.
+   - `scan-history.tsv` → same fuzzy match against company + title columns (not just URL). A role reposted on a new URL but with the same company and similar title is a duplicate.
+   - `pipeline.md` → same fuzzy match against company + title in pending items that include metadata (format: `- [ ] {url} | {company} | {title}`)
+
+   **Fuzzy match rules:**
+   - Normalize company: `company.toLowerCase().replace(/[^a-z0-9]/g, '')`
+   - Fuzzy role match: split both titles into words > 3 chars, match if 2+ words overlap (substring match, case-insensitive). E.g., "Senior AI Engineer" and "Staff AI Engineer" share "engineer" — only 1 overlap, not a match. But "AI Platform Engineer" and "AI Platform Eng" share "platform" + partial "engineer" — match.
+   - When a fuzzy match is found but the URL is new, log it as `skipped_repost` (not `skipped_dup`) with a note referencing the original entry number.
+
+8. **For each new offer that passes filters**:
+   a. Add to `pipeline.md` section "Pending": `- [ ] {url} | {company} | {title}`
+   b. Record in `scan-history.tsv`: `{url}\t{date}\t{query_name}\t{title}\t{company}\tadded`
+
+9. **Offers filtered by title**: record in `scan-history.tsv` with status `skipped_title`
+10. **Duplicate offers (URL-exact)**: record with status `skipped_dup`
+11. **Duplicate offers (fuzzy repost)**: record with status `skipped_repost` and note `repost of #{original_entry_num}`
+
+## Extract Title And Company From WebSearch Results
+
+WebSearch results come in the format: `"Job Title @ Company"` or `"Job Title | Company"` or `"Job Title — Company"`.
+
+Extraction patterns by portal:
+- **Ashby**: `"Senior AI PM (Remote) @ EverAI"` → title: `Senior AI PM`, company: `EverAI`
+- **Greenhouse**: `"AI Engineer at Anthropic"` → title: `AI Engineer`, company: `Anthropic`
+- **Lever**: `"Product Manager - AI @ Temporal"` → title: `Product Manager - AI`, company: `Temporal`
+
+Generic regex: `(.+?)(?:\s*[@|—–-]\s*|\s+at\s+)(.+?)$`
+
+## Resolve Private URLs
+
+If a publicly inaccessible URL is found:
+1. Save the JD to `jds/{company}-{role-slug}.md`
+2. Add to pipeline.md as: `- [ ] local:jds/{company}-{role-slug}.md | {company} | {title}`
+
+## Scan History
+
+`data/scan-history.tsv` tracks ALL seen URLs:
+
+```
+url	first_seen	portal	title	company	status
+https://...	2026-02-10	Ashby — AI PM	PM AI	Acme	added
+https://...	2026-02-10	Greenhouse — SA	Junior Dev	BigCo	skipped_title
+https://...	2026-02-10	Ashby — AI PM	SA AI	OldCo	skipped_dup
+```
+
+## Output Summary
+
+```
+Portal Scan — {YYYY-MM-DD}
+━━━━━━━━━━━━━━━━━━━━━━━━━━
+Queries executed: N
+Offers found: N total
+Filtered by title: N relevant
+Duplicates: N (already evaluated or in pipeline)
+New added to pipeline.md: N
+
+  + {company} | {title} | {query_name}
+  ...
+
+→ Run /job-forge pipeline to evaluate the new offers.
+```
+
+## Update careers_url
+
+Each company in `tracked_companies` MUST have a `careers_url` — the direct URL to its job listings page. The stored URL avoids searching for it every time.
+
+**Known patterns by platform:**
+- **Ashby:** `https://jobs.ashbyhq.com/{slug}`
+- **Greenhouse:** `https://job-boards.greenhouse.io/{slug}` or `https://job-boards.eu.greenhouse.io/{slug}`
+- **Lever:** `https://jobs.lever.co/{slug}`
+- **Custom:** The company's own URL (e.g., `https://openai.com/careers`)
+
+**If `careers_url` doesn't exist** for a company:
+1. Try the pattern for its known platform
+2. If that fails, do a quick WebSearch: `"{company}" careers jobs`
+3. Navigate with Geometra (`geometra_connect`) to confirm it works
+4. **Save the found URL in portals.yml** for future scans
+
+**If `careers_url` returns 404 or redirect:**
+1. Note in the output summary
+2. Try scan_query as fallback
+3. Flag for manual update
+
+## Update portals.yml
+
+- **ALWAYS save `careers_url`** when adding a new company
+- Add new queries as interesting portals or roles are discovered
+- Disable queries with `enabled: false` if they generate too much noise
+- Adjust filtering keywords as target roles evolve
+- Add companies to `tracked_companies` when you want to follow them closely
+- Verify `careers_url` periodically — companies change ATS platforms

package/modes/tracker.md

@@ -0,0 +1,31 @@
+# Mode: tracker — Application Tracker
+
+Reads and displays the application tracker: day-based files in `data/applications/` (format: `YYYY-MM-DD.md`).
+
+**Tracker format:**
+```markdown
+| # | Date | Company | Role | Score | Status | PDF | Report | Notes |
+|---|------|---------|------|-------|--------|-----|--------|-------|
+```
+
+Possible states (canonical, per `templates/states.yml`):
+
+`Evaluated` → `Applied` → `Contacted` → `Responded` → `Interview` → `Offer` / `Rejected` / `Discarded` / `SKIP`
+
+- `Applied` = the candidate submitted their application
+- `Contacted` = the candidate proactively reached out to someone at the company (outbound, e.g., LinkedIn power move via `/job-forge contact`)
+- `Responded` = a recruiter/company contacted back and the candidate responded (inbound)
+
+If the user asks to update a status, edit the corresponding row in the day file where the entry exists.
+
+Also display statistics:
+- Total applications
+- By status
+- Average score
+- % with generated PDF
+- % with generated report
+
+If any entries look overdue for follow-up (Applied 7+ days ago, Contacted 5+ days ago, Interviewed with no update 7+ days), mention it:
+> "3 entries may need follow-up. Run `/job-forge followup` for details."
+
+The followup reminder above is a passive hint — it does NOT change tracker behavior or output format.

package/modes/training.md

@@ -0,0 +1,27 @@
+# Mode: training — Training Evaluation
+
+For each course/cert the candidate asks about, evaluate across 6 dimensions:
+
+| Dimension | What it evaluates |
+|-----------|-------------------|
+| North Star alignment | Does it move closer to or away from the goal? |
+| Recruiter signal | What do HMs think when they see this on a CV? |
+| Time and effort | Weeks x hours/week |
+| Opportunity cost | What can't they do during that time? |
+| Risks | Outdated content? Weak brand? Too basic? |
+| Portfolio deliverable | Does it produce a demonstrable artifact? |
+
+## Return One Of These Verdicts
+
+- **DO IT** → 4-12 week plan with weekly deliverables and scoreboard
+- **DON'T DO IT** → better alternative with justification
+- **DO IT WITH TIMEBOX** (max X weeks) → condensed plan, essentials only
+
+## Apply This Priority Order
+
+Training that improves credibility in "production-grade AI":
+1. Evals and LLM testing
+2. Observability and monitoring
+3. Cost/reliability trade-offs
+4. AI governance and safety
+5. Enterprise AI architecture

package/normalize-statuses.mjs

@@ -0,0 +1,152 @@
+#!/usr/bin/env node
+/**
+ * normalize-statuses.mjs — Clean non-canonical states in the application tracker
+ *
+ * Supports both layouts:
+ *   - Day-based: data/applications/YYYY-MM-DD.md (preferred)
+ *   - Single-file: data/applications.md or applications.md (legacy)
+ *
+ * Maps all non-canonical statuses to canonical ones per templates/states.yml:
+ *   Evaluated, Applied, Responded, Contacted, Interview, Offer, Rejected, Discarded, SKIP
+ *
+ * Also strips markdown bold (**) and dates from the status field,
+ * moving DUPLICADO info to the notes column.
+ *
+ * Run: node normalize-statuses.mjs [--dry-run]   (from repo root)
+ */
+
+import { readFileSync, writeFileSync, copyFileSync, existsSync } from 'fs';
+import { join, relative, dirname } from 'path';
+import { fileURLToPath } from 'url';
+import {
+  PROJECT_DIR, DATA_APPS_DIR, DATA_APPS_FILE, ROOT_APPS_FILE,
+  usesDayFiles, ensureDayDir, parseAppLine, formatAppLine,
+  readAllEntries, writeToDayFiles, listDayFiles,
+} from './tracker-lib.mjs';
+
+const DRY_RUN = process.argv.includes('--dry-run');
+
+if (process.argv.includes('--help') || process.argv.includes('-h')) {
+  console.log(`normalize-statuses.mjs — map tracker status column to canonical labels
+
+Supports day-based (data/applications/YYYY-MM-DD.md) and single-file layouts.
+Uses templates/states.yml display labels when present. Strips markdown bold
+and dates from the status field; moves duplicate/repost markers into notes
+where applicable.
+
+Usage:
+  node normalize-statuses.mjs [--dry-run]
+  npm run normalize [-- --dry-run]
+
+Exits successfully when no tracker entries exist (nothing to do).
+Creates a .bak copy next to the tracker before writing (single-file mode).
+
+Run from the repository root.`);
+  process.exit(0);
+}
+
+function normalizeStatus(raw) {
+  let s = raw.replace(/\*\*/g, '').trim();
+  const lower = s.toLowerCase();
+
+  if (/^dup(licate)?/i.test(s)) {
+    return { status: 'Discarded', moveToNotes: raw.trim() };
+  }
+
+  if (/^contacted$/i.test(s)) return { status: 'Contacted' };
+
+  if (/^hold$/i.test(s)) return { status: 'Evaluated' };
+
+  if (/^repost/i.test(s)) return { status: 'Discarded', moveToNotes: raw.trim() };
+
+  if (s === '—' || s === '-' || s === '') return { status: 'Discarded' };
+
+  const canonical = [
+    'Evaluated', 'Applied', 'Contacted', 'Responded', 'Interview',
+    'Offer', 'Rejected', 'Discarded', 'SKIP',
+  ];
+  for (const c of canonical) {
+    if (lower === c.toLowerCase()) return { status: c };
+  }
+
+  if (['applied', 'sent'].includes(lower)) return { status: 'Applied' };
+  if (['skip'].includes(lower)) return { status: 'SKIP' };
+
+  return { status: null, unknown: true };
+}
+
+// Read entries
+const { entries, source } = readAllEntries();
+
+if (entries.length === 0) {
+  console.log('No tracker entries found. Nothing to normalize.');
+  process.exit(0);
+}
+
+let changes = 0;
+let unknowns = [];
+const updated = entries.map(app => {
+  const result = normalizeStatus(app.status);
+
+  if (result.unknown) {
+    unknowns.push({ num: app.num, rawStatus: app.status });
+    return app;
+  }
+
+  if (result.status === app.status) return app;
+
+  changes++;
+  console.log(`#${app.num}: "${app.status}" → "${result.status}"`);
+
+  let notes = app.notes || '';
+  if (result.moveToNotes && !notes.includes(result.moveToNotes)) {
+    notes = result.moveToNotes + (notes ? '. ' + notes : '');
+  }
+
+  // Also strip bold from score
+  const score = app.score ? app.score.replace(/\*\*/g, '') : app.score;
+
+  return { ...app, status: result.status, notes, score };
+});
+
+if (unknowns.length > 0) {
+  console.log(`\n⚠️  ${unknowns.length} unknown statuses:`);
+  for (const u of unknowns) {
+    console.log(`  #${u.num}: "${u.rawStatus}"`);
+  }
+}
+
+console.log(`\n📊 ${changes} statuses normalized`);
+
+if (!DRY_RUN && changes > 0) {
+  if (source === 'day') {
+    writeToDayFiles(updated);
+    console.log('✅ Written to day files');
+  } else {
+    const APPS_FILE = existsSync(DATA_APPS_FILE) ? DATA_APPS_FILE : ROOT_APPS_FILE;
+    const appsDisplay = relative(PROJECT_DIR, APPS_FILE).replace(/\\/g, '/');
+    copyFileSync(APPS_FILE, APPS_FILE + '.bak');
+    // Rewrite single-file
+    const filePath = APPS_FILE;
+    const content = readFileSync(filePath, 'utf-8');
+    const lines = content.split('\n');
+    const updatedLines = [];
+    for (const line of lines) {
+      const app = parseAppLine(line);
+      if (app) {
+        const newApp = updated.find(u => u.num === app.num);
+        if (newApp) {
+          updatedLines.push(formatAppLine(newApp));
+          continue;
+        }
+      }
+      updatedLines.push(line);
+    }
+    writeFileSync(filePath, updatedLines.join('\n'));
+    console.log(`✅ Written to ${appsDisplay} (backup: ${appsDisplay}.bak)`);
+  }
+} else if (DRY_RUN) {
+  console.log('(dry-run — no changes written)');
+} else {
+  console.log('✅ No changes needed');
+}

package/opencode.json

@@ -0,0 +1,28 @@
+{
+  "$schema": "https://opencode.ai/config.json",
+  "mcp": {
+    "geometra": {
+      "type": "local",
+      "command": [
+        "npx",
+        "-y",
+        "@geometra/mcp"
+      ],
+      "environment": {}
+    },
+    "gmail": {
+      "type": "local",
+      "command": [
+        "npx",
+        "-y",
+        "@razroo/gmail-mcp"
+      ],
+      "environment": {
+        "DISABLE_HTTP": "true"
+      }
+    }
+  },
+  "instructions": [
+    "templates/states.yml"
+  ]
+}

package/package.json

@@ -0,0 +1,78 @@
+{
+  "name": "job-forge",
+  "version": "2.0.0",
+  "description": "AI-powered job search pipeline built on opencode",
+  "type": "module",
+  "bin": {
+    "job-forge": "bin/job-forge.mjs",
+    "create-job-forge": "bin/create-job-forge.mjs"
+  },
+  "scripts": {
+    "verify": "node verify-pipeline.mjs",
+    "build:dashboard": "cd dashboard && go build .",
+    "normalize": "node normalize-statuses.mjs",
+    "dedup": "node dedup-tracker.mjs",
+    "merge": "node merge-tracker.mjs",
+    "pdf": "node generate-pdf.mjs",
+    "sync-check": "node cv-sync-check.mjs",
+    "tokens": "node scripts/token-usage-report.mjs",
+    "tokens:today": "node scripts/token-usage-report.mjs --days 1",
+    "tokens:log": "node scripts/token-usage-report.mjs --days 1 --append",
+    "build:config": "iso-harness build --source iso --out .",
+    "prepack": "iso-harness build --source iso --out .",
+    "release:check-source": "node ./scripts/release/check-source.mjs",
+    "postinstall": "node bin/sync.mjs"
+  },
+  "files": [
+    "bin/",
+    "iso/",
+    ".cursor/mcp.json",
+    ".cursor/rules/",
+    ".opencode/",
+    ".codex/",
+    ".mcp.json",
+    "CLAUDE.md",
+    "AGENTS.md",
+    "opencode.json",
+    "modes/",
+    "templates/",
+    "config/profile.example.yml",
+    "fonts/",
+    "scripts/",
+    "batch/batch-prompt.md",
+    "batch/batch-runner.sh",
+    "batch/README.md",
+    "docs/",
+    "tracker-lib.mjs",
+    "merge-tracker.mjs",
+    "dedup-tracker.mjs",
+    "verify-pipeline.mjs",
+    "normalize-statuses.mjs",
+    "generate-pdf.mjs",
+    "cv-sync-check.mjs",
+    "README.md",
+    "LICENSE"
+  ],
+  "keywords": [
+    "ai",
+    "job-search",
+    "opencode",
+    "career",
+    "automation"
+  ],
+  "author": "Charlie Greenman",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/razroo/JobForge"
+  },
+  "license": "MIT",
+  "engines": {
+    "node": ">=18"
+  },
+  "dependencies": {
+    "playwright": "^1.58.1"
+  },
+  "devDependencies": {
+    "@razroo/iso-harness": "^0.1.3"
+  }
+}