feat: add jobbasel.ch and englishjobsearch.ch job sources

jobbasel.ch's search route (/job/suche-{slug}) is server-rendered with
a window.__PRELOADED_STATE__ JSON blob, no headless browser needed.
englishjobsearch.ch works the same way via its ?format=markdown param.
Both swept with the same fixed keyword list job-room.ch already uses.

basel-jobs.ch was evaluated too but left out: it's DataDome-protected,
which blocks plain curl outright, and this project's fetchers are
intentionally browser-free.

While wiring these in, found that isEnglishDescription (now
isViableDescription) only ever ran the DE/FR word-count check against
inline descriptions, never requiresGerman() - so an English-written
posting demanding fluent German (e.g. PRODYNA's "you can communicate
fluently in both German and English") sailed straight through. That
gap existed in itjobs.ch and job-room.ch too, not just the two new
sources, so the fix applies to all four.

CLAUDE.md

@@ -20,7 +20,7 @@ Temp scripts: `bun run scripts/<name>-temp.ts` — never edit `config.ts` for on
 ```
 config.ts          # search keywords and areas
 fetch.ts           # scraper, HTML parser, all filters
-fetch-other.ts     # extra sources (datacareer.ch, stellen-basel.ch, itjobs.ch, job-room.ch, Basel intl orgs)
+fetch-other.ts     # extra sources (datacareer.ch, stellen-basel.ch, itjobs.ch, job-room.ch, Basel intl orgs, jobbasel.ch, englishjobsearch.ch)
 index.ts           # orchestrator: runs searches, writes CSV
 recheck.ts         # raw-mode fetch, lists filtered-out jobs
 daily.ts           # digest: fetch → match against profile.ts → send to Telegram
@@ -42,8 +42,11 @@ plans/             # LinkedIn post drafts
   - Also exports `PLATFORM_KEYWORDS` (architect/SRE/infra/security/MLOps), `RESEARCH_KEYWORDS` (research & scientific software, HPC, data platform) and `FIELD_KEYWORDS` (forward-deployed, customer/sales/integration, founding/product). Kept as separate OR-strings rather than folded into `KEYWORDS` — LinkedIn ranks a focused query far better than one 50-term disjunction. The `sweep()` helper generates one search per area per family; Research skips Bern, Platform and Field add a Switzerland-wide remote pass.
   - `RESEARCH_KEYWORDS` deliberately contains no bioinformatics/computational-biology terms — see `BIO_DOMAIN_RE`.
 - **`fetch.ts`** — HTTP scraping, HTML parsing, and all filters. Exports `fetchJobs(config, { raw? })` and `isEnglishContent(id)`.
-- **`fetch-other.ts`** — scrapes extra sources (datacareer.ch, stellen-basel.ch, itjobs.ch, job-room.ch, Basel international orgs). Imported by `index.ts`.
+- **`fetch-other.ts`** — scrapes extra sources (datacareer.ch, stellen-basel.ch, itjobs.ch, job-room.ch, Basel international orgs, jobbasel.ch, englishjobsearch.ch). Imported by `index.ts`.
   - **datacareer.ch** currently returns **403** on every page — the source is dead and yields nothing. Needs new headers or a different entry point.
+  - **jobbasel.ch** looks like a JS-only React/Vite SPA (no jobs in the homepage's raw HTML), but its search route is server-rendered: `GET /job/suche-{slug}` returns full HTML with a `window.__PRELOADED_STATE__` blob containing the matched vacancies as JSON — no headless browser needed. Pagination is a URL suffix (`-seite-{n}`); a bare `?page=` query param is silently ignored. Swept with the same fixed keyword list as job-room.ch (no keyword-union endpoint exists) and deduped by vacancy id. Each item carries its full German description, so the DE/FR check runs in-process.
+  - **englishjobsearch.ch** also looks JS-only at first glance, but any page accepts `?format=markdown` and returns clean server-rendered markdown with full job cards — title, company, city, date, description. Hit the canton pages directly (`/in/canton-basel-stadt`, `/in/canton-basel-landschaft`) rather than the `/in/basel/` landing page, which only lists canton counts. Every job link is a `/clickout/<hash>` tracking redirect rather than the real posting URL — there's no way to recover the underlying employer link without following the redirect, so the clickout link is stored as-is.
+  - **basel-jobs.ch** was evaluated but is **not integrated**: it's protected by DataDome (not basic Cloudflare), which blocks plain `curl` outright. A real Chrome session gets through, confirming the content is genuine, but that only proves the block isn't automatable within this project's curl-only architecture — `bun daily`/`bun fetch` run headless via launchd with no browser available, and the session cookie that lets a real browser through can't be extracted for replay (browser tooling blocks cookie access). Would need persistent Playwright + stealth infra to chase further, which isn't worth it for what's likely a small regional board already covered by jobbasel.ch and stellen-basel.ch.
   - **Basel international orgs** (`fetchBaselIntlOrgs`) sweeps three boards under one label, each only a handful of postings:
     - **BIS** — the career site is JS-rendered with no usable API, so we read the UNjobs mirror at `unjobs.org/offices/bis_182320855689338880`. Titles arrive as `Role, City, Country` and get split back apart. BIS also cycles Senior Software / Data / Data & Analytics Engineer roles through this board.
     - **Swiss TPH** — Umantis at `jobs.swisstph.ch/Jobs/All`. Site chrome is German even on English postings, so `isEnglishUrl` would reject everything; `contentChecked: true` and the title filters decide instead.

fetch-other.ts

@@ -1,4 +1,4 @@
-import { NON_ENGLISH_RE, JUNIOR_TITLE_RE, SOFTWARE_TITLE_RE } from "./fetch.ts";
+import { NON_ENGLISH_RE, JUNIOR_TITLE_RE, SOFTWARE_TITLE_RE, requiresGerman } from "./fetch.ts";
 import type { Job } from "./fetch.ts";
 
 const HEADERS = {
@@ -34,10 +34,22 @@ const MS_STACK_RE = /\.net\b|\bc#|sharepoint|dynamics 365|power platform|\bsccm\
 const DE_FR_WORDS_RE =
   /\b(und|oder|wir|Sie|Ihre|Ihrer|Kenntnisse|Erfahrung|Aufgaben|Anforderungen|Stelle|Unternehmen|Mitarbeiter|Bereich|nous|vous|votre|pour|avec|des|les)\b/g;
 
-function isEnglishDescription(description?: string): boolean {
-  if (!description) return true; // nothing to judge — let the URL check in index.ts decide
+// English-language AND not demanding German above B1 — mirrors isViableContent/
+// isViableUrl in fetch.ts, but against a description we already have in hand
+// instead of a page fetch. The two checks catch different things: the word
+// count drops postings *written* in German/French; requiresGerman() drops
+// postings *written in English* that still demand German above B1 (e.g.
+// PRODYNA: "you can communicate fluently in both German and English" — clean
+// English text, would sail through the word count untouched). Skipping the
+// requiresGerman() half was a real gap in every source below until 2026-08-30
+// — it let German-fluency requirements slip past whenever an item's inline
+// description read as English.
+function isViableDescription(description?: string): { ok: boolean; reason?: string } {
+  if (!description) return { ok: true }; // nothing to judge — let the URL check in index.ts decide
   const text = description.replace(/<[^>]+>/g, " ").replace(/&[a-z]+;/g, " ");
-  return (text.match(DE_FR_WORDS_RE) ?? []).length < 6;
+  if ((text.match(DE_FR_WORDS_RE) ?? []).length >= 6) return { ok: false, reason: "non-English" };
+  if (requiresGerman(text)) return { ok: false, reason: "requires German > B1" };
+  return { ok: true };
 }
 
 // datacareer.ch — HTML listing at /jobs/?city=Basel
@@ -152,7 +164,7 @@ export async function fetchItJobsCh(): Promise<Job[]> {
       if (seen.has(id)) continue;
       seen.add(id);
 
-      if (!isEnglishDescription(item.description)) { langDropped++; continue; }
+      if (!isViableDescription(item.description).ok) { langDropped++; continue; }
 
       allJobs.push({
         id,
@@ -165,7 +177,7 @@ export async function fetchItJobsCh(): Promise<Job[]> {
         link: `https://www.itjobs.ch${item.job_details_path}`,
         search: "itjobs.ch",
         source: "itjobs.ch",
-        contentChecked: true, // isEnglishDescription already ran on the real description
+        contentChecked: true, // isViableDescription already ran on the real description
       });
       newOnPage++;
     }
@@ -274,12 +286,12 @@ export async function fetchJobRoom(): Promise<Job[]> {
     if (MS_STACK_RE.test(title)) continue;
     if (PHARMA_ROLE_RE.test(title)) continue;
 
-    // A handful of ads ship without a description. isEnglishDescription passes
+    // A handful of ads ship without a description. isViableDescription passes
     // those through by design, so leave contentChecked unset and let index.ts
     // fall back to fetching the page.
     const body = desc.description ?? "";
     const hasBody = body.trim().length > 40;
-    if (hasBody && !isEnglishDescription(body)) { langDropped++; continue; }
+    if (hasBody && !isViableDescription(body).ok) { langDropped++; continue; }
 
     const city = content?.location?.city ?? content?.company?.city ?? "";
     const canton = content?.location?.cantonCode ?? "";
@@ -456,6 +468,193 @@ export async function fetchBaselIntlOrgs(): Promise<Job[]> {
   return jobs;
 }
 
+// jobbasel.ch — the page shell is a React/Vite SPA (no jobs in the raw HTML
+// on the homepage), but the search results route is server-rendered: fetching
+// `/job/suche-{slug}` returns full HTML with a `window.__PRELOADED_STATE__`
+// blob already containing the matched vacancies as JSON. No headless browser
+// needed. Pagination is a URL suffix (`-seite-{n}`), not a query param — a
+// bare `?page=2` is silently ignored by the SPA route.
+//
+// The board (Basel-Stadt + Basel-Landschaft, ~4000 openings) has no keyword
+// union endpoint, so — like job-room.ch — we sweep a fixed keyword list and
+// dedupe by vacancy id. Each item carries its full German description
+// (activity/requirements/offer), so the DE/FR check runs here instead of
+// costing a page fetch per job in index.ts.
+const JOBBASEL_KEYWORDS = [
+  "software", "engineer", "developer", "devops", "data", "python",
+  "cloud", "platform", "typescript", "backend", "full stack", "ai",
+];
+
+interface JobBaselItem {
+  id: number;
+  title: string;
+  activity?: string;
+  requirements?: string;
+  offer?: string;
+  workplaceCity?: string;
+  urlApplication?: string;
+  dateFirstPublished?: string;
+  company?: { name?: string };
+}
+
+// Balanced-brace extraction — the blob can itself contain "};" inside string
+// values (job descriptions), so a regex cutoff would truncate it.
+function extractPreloadedState(html: string): any | null {
+  const marker = "window.__PRELOADED_STATE__";
+  const markerIdx = html.indexOf(marker);
+  if (markerIdx === -1) return null;
+  const start = html.indexOf("{", markerIdx);
+  if (start === -1) return null;
+
+  let depth = 0, inStr = false, esc = false, i = start;
+  for (; i < html.length; i++) {
+    const c = html[i];
+    if (inStr) {
+      if (esc) esc = false;
+      else if (c === "\\") esc = true;
+      else if (c === '"') inStr = false;
+    } else {
+      if (c === '"') inStr = true;
+      else if (c === "{") depth++;
+      else if (c === "}") { depth--; if (depth === 0) { i++; break; } }
+    }
+  }
+
+  try { return JSON.parse(html.slice(start, i)); } catch { return null; }
+}
+
+export async function fetchJobBasel(): Promise<Job[]> {
+  const items = new Map<number, JobBaselItem>();
+
+  for (const keyword of JOBBASEL_KEYWORDS) {
+    process.stdout.write(`  ${keyword} … `);
+    const slug = keyword.toLowerCase().replace(/\s+/g, "-");
+    let total = 0;
+
+    for (let page = 1; page <= 5; page++) {
+      if (page > 1) await Bun.sleep(300);
+      const suffix = page === 1 ? "" : `-seite-${page}`;
+      const res = await fetch(
+        `https://www.jobbasel.ch/job/suche-${slug}${suffix}`,
+        { headers: HEADERS }
+      );
+      if (!res.ok) { console.log(`failed (${res.status})`); break; }
+
+      const state = extractPreloadedState(await res.text());
+      const pageItems: JobBaselItem[] = state?.vacancySearch?.data?.items ?? [];
+      if (!pageItems.length) break;
+
+      for (const item of pageItems) items.set(item.id, item);
+      total += pageItems.length;
+      if (pageItems.length < 10) break; // last page (pageSize is 10)
+    }
+    console.log(`${total} ads (${items.size} unique)`);
+    await Bun.sleep(300);
+  }
+
+  const jobs: Job[] = [];
+  let langDropped = 0;
+  for (const item of items.values()) {
+    const title = decodeHtml(item.title.replace(/<[^>]+>/g, ""));
+    if (!applyFilters(title)) continue;
+    if (MS_STACK_RE.test(title)) continue;
+
+    const body = [item.activity, item.requirements, item.offer]
+      .filter(Boolean)
+      .join(" ")
+      .replace(/<[^>]+>/g, " ");
+    const hasBody = body.trim().length > 40;
+    if (hasBody && !isViableDescription(body).ok) { langDropped++; continue; }
+
+    jobs.push({
+      id: `jb-${item.id}`,
+      title,
+      company: item.company?.name ?? "",
+      location: item.workplaceCity ? `${item.workplaceCity}, Switzerland` : "Basel, Switzerland",
+      postedAt: item.dateFirstPublished?.slice(0, 10) ?? "",
+      link: item.urlApplication || `https://www.jobbasel.ch/job/${item.id}`,
+      search: "jobbasel.ch",
+      source: "jobbasel.ch",
+      contentChecked: hasBody,
+    });
+  }
+
+  console.log(`  ${items.size} ads scanned, ${jobs.length} kept, ${langDropped} non-English`);
+  return jobs;
+}
+
+// englishjobsearch.ch — plain SSR HTML, no JS needed, but reached through a
+// non-obvious lever: any page on the site accepts `?format=markdown` and
+// returns the same content pre-converted to clean markdown text (job title,
+// company, city, date, short description all present). Without that param
+// the canton pages ship the job cards as plain text nodes too — no JSON blob,
+// no client-side rendering — so this works with a bare `curl`-equivalent
+// fetch either way; `?format=markdown` just makes parsing trivial.
+//
+// Two catches: (1) the directory root groups by canton, so we hit the
+// Basel-Stadt and Basel-Landschaft canton pages directly rather than the
+// generic `/in/basel/` landing page, which only lists canton counts.
+// (2) every job link is a `/clickout/<hash>` tracking redirect, not the real
+// posting URL — there is no way to recover the underlying employer link
+// without following the redirect, so the clickout link is stored as-is (it
+// still resolves for a human clicking it, and content-checking follows it
+// server-side the same as any other link).
+const ENGLISHJOBSEARCH_CANTONS = ["canton-basel-stadt", "canton-basel-landschaft"];
+
+export async function fetchEnglishJobSearch(): Promise<Job[]> {
+  const jobs: Job[] = [];
+  const seen = new Set<string>();
+
+  for (const canton of ENGLISHJOBSEARCH_CANTONS) {
+    process.stdout.write(`  ${canton} … `);
+    const res = await fetch(
+      `https://englishjobsearch.ch/in/${canton}?format=markdown`,
+      { headers: HEADERS }
+    );
+    if (!res.ok) { console.log(`failed (${res.status})`); continue; }
+
+    const md = await res.text();
+    // Each posting: "### Title](/clickout/<hash>?...)" heading, then on
+    // separate lines "* Company", "* City", "* Date", then a description
+    // paragraph before the next "### " heading or "report probem" footer.
+    const cards = [...md.matchAll(
+      /### ([^\]]+)\]\((\/clickout\/[^)]+)\)\n\n\* ([^\n]+)\n\* ([^\n]+)\n\* ([^\n]+)\n\n([\s\S]*?)\n\nreport probem/g
+    )];
+
+    let kept = 0;
+    for (const [, rawTitle, clickout, company, city, dateStr, desc] of cards) {
+      const title = rawTitle.trim();
+      if (!applyFilters(title)) continue;
+      if (MS_STACK_RE.test(title)) continue;
+
+      const id = clickout.match(/\/clickout\/([a-f0-9]+)/)?.[1] ?? clickout;
+      if (seen.has(id)) continue;
+      seen.add(id);
+
+      const description = desc.trim();
+      const hasBody = description.length > 40;
+      if (hasBody && !isViableDescription(description).ok) continue;
+
+      jobs.push({
+        id: `ejs-${id}`,
+        title,
+        company: company.trim(),
+        location: `${city.trim()}, Switzerland`,
+        postedAt: "", // "August 30" style, no year given — not worth mis-parsing
+        link: `https://englishjobsearch.ch${clickout}`,
+        search: "englishjobsearch.ch",
+        source: "englishjobsearch.ch",
+        contentChecked: hasBody,
+      });
+      kept++;
+    }
+    console.log(`${cards.length} ads, ${kept} kept`);
+    await Bun.sleep(300);
+  }
+
+  return jobs;
+}
+
 // stellen-basel.ch — WordPress REST API
 export async function fetchStellenBasel(): Promise<Job[]> {
   const allJobs: Job[] = [];

fetch.ts

@@ -158,7 +158,7 @@ const GERMAN_OPTIONAL_RE =
 
 /** Within a mention's window: signals it is genuinely required at B2+. */
 const GERMAN_REQUIRED_RE =
-  /\b(?:required|require[sd]?|must|mandatory|essential|necessary|fluent|fluency|proficien\w*|excellent|strong|very\s+good|good\s+(?:knowledge|command|level)|native|mother\s*tongue|business\s+level|negotiat\w*|B2|C1|C2|verhandlungssicher\w*|fliessend|fließend|muttersprach\w*|sehr\s+gute?|gute?\s+Deutschkenntnisse|voraussetzung|zwingend)\b/i;
+  /\b(?:required|require[sd]?|must|mandatory|essential|necessary|fluen\w*|proficien\w*|excellent|strong|very\s+good|good\s+(?:knowledge|command|level)|native|mother\s*tongue|business\s+level|negotiat\w*|B2|C1|C2|verhandlungssicher\w*|fliessend|fließend|muttersprach\w*|sehr\s+gute?|gute?\s+Deutschkenntnisse|voraussetzung|zwingend)\b/i;
 
 /** Title-level marker — no description fetch needed. */
 export const GERMAN_TITLE_RE =