Arbeitnow reads what Work reads (pages 1-3, a plain-text summary); a capped HTML summary keeps no half tag (#91)

ddiff --git a/seeds/sources.json b/seeds/sources.json
iindex 5bab631..f500030 100644
---- a/seeds/sources.json
++++ b/seeds/sources.json
@@@ -1753,6 +1753,43 @@
              "identity": "slug",
              "url": "url",
              "title": "title",
++            "summary": "description",
++            "published_at_epoch": "created_at"
++          },
++          "extra": {
++            "company": "company_name",
++            "location": "location",
++            "remote": "remote"
++          }
++        },
++        {
++          "url": "https://www.arbeitnow.com/api/job-board-api?page=2",
++          "format": "json",
++          "items_path": "data",
++          "item_kind": "job",
++          "fields": {
++            "identity": "slug",
++            "url": "url",
++            "title": "title",
++            "summary": "description",
++            "published_at_epoch": "created_at"
++          },
++          "extra": {
++            "company": "company_name",
++            "location": "location",
++            "remote": "remote"
++          }
++        },
++        {
++          "url": "https://www.arbeitnow.com/api/job-board-api?page=3",
++          "format": "json",
++          "items_path": "data",
++          "item_kind": "job",
++          "fields": {
++            "identity": "slug",
++            "url": "url",
++            "title": "title",
++            "summary": "description",
              "published_at_epoch": "created_at"
            },
            "extra": {
@@@ -1763,7 +1800,7 @@
          }
        ],
        "robots_note": "robots.txt checked 2026-10-22: User-agent * with an empty Disallow (only ?__hstc tracking URLs excluded). The fetcher re-checks it on every run.",
--      "terms_note": "Arbeitnow publishes its job board API free of charge (the terms OpenVibe.Work records today): attribute Arbeitnow by name and link every listing to its page on arbeitnow.com. The description HTML is deliberately not mapped. The source OpenVibe.Work fetches itself today; Work moves to it here. Re-verify before enabling.",
++      "terms_note": "Arbeitnow publishes its job board API free of charge (the terms OpenVibe.Work records today): attribute Arbeitnow by name and link every listing to its page on arbeitnow.com. The first three pages are read, as Work read them. The description is kept only as a short plain-text excerpt (the summary, as Work shows it); every listing links to arbeitnow.com for the full text. A job the board stops listing is retired after a week. Work moves to this source from its own fetch.",
        "license_note": null,
        "min_interval_ms": 60000,
        "poll_interval_sec": 21600,
ddiff --git a/server/adapters/util.js b/server/adapters/util.js
iindex baaee4a..712678a 100644
---- a/server/adapters/util.js
++++ b/server/adapters/util.js
@@@ -59,7 +59,8 @@ function toText(value, max = 1000) {
      s = decodeEntities(String(s))
          .replace(/<(script|style|noscript)[\s\S]*?<\/\1>/gi, ' ')
          .replace(/<br\s*\/?>|<\/p>|<\/li>|<\/h\d>/gi, ' ')
--        .replace(/<[^>]*>/g, ' ');
++        .replace(/<[^>]*>/g, ' ')
++        .replace(/<[^>]*$/, ' ');   // a tag cut off where a capped value ended
      s = decodeEntities(s).replace(/[\u0000-\u0008\u000B\u000C\u000E-\u001F\u007F]/g, '').replace(/\s+/g, ' ').trim();
      if (!s) return null;
      return s.length > max ? `${s.slice(0, max - 1).trimEnd()}…` : s;
ddiff --git a/test/adapters.test.js b/test/adapters.test.js
iindex 06adb82..4354716 100644
---- a/test/adapters.test.js
++++ b/test/adapters.test.js
@@@ -196,4 +196,13 @@ t('helpers: charset decoding, date and URL strictness', () => {
      assert.strictEqual(canonicalUrl('HTTPS://Example.COM:443/a?gclid=1&b=2#x'), 'https://example.com/a?b=2');
  });
  
++t('an api summary from HTML longer than the 500-character cap keeps no half tag (Arbeitnow descriptions)', () => {
++    const description = `<p><strong>About us</strong></p><p>${'We build things. '.repeat(26)}</p><p><a href="https://example.org/apply">Apply</a></p>`;
++    const body = JSON.stringify({ data: [{ slug: 'dev-1', url: 'https://www.arbeitnow.com/jobs/dev-1', title: 'Developer', description, created_at: 1761177600 }] });
++    const endpoint = { format: 'json', items_path: 'data', item_kind: 'job', fields: { identity: 'slug', url: 'url', title: 'title', summary: 'description', published_at_epoch: 'created_at' } };
++    const [item] = api.parse(body, { url: 'https://www.arbeitnow.com/api/job-board-api', endpoint }).items;
++    assert.match(item.summary, /^About us We build things\./);
++    assert.ok(!/[<>]/.test(item.summary), `no markup left: ${item.summary.slice(-60)}`);
++});
++
  t.run();