jackwener/OpenCLI · error · CommandExecutionError

NotebookLM home-link probe returned an untrusted or mismatch

Error message

NotebookLM home-link probe returned an untrusted or mismatched notebook URL

What it means

After a home-link row passes shape validation, its url/id pair is normalized via normalizeNotebooklmNotebookUrl, which only accepts trusted NotebookLM notebook URLs with a matching id. If normalization yields nothing, the row's link is untrusted (wrong host/path) or inconsistent with its id, so the CLI throws rather than emitting a bad notebook link.

Source

Thrown at clis/notebooklm/utils.js:801

        title,
        url: href,
        source: 'home-links',
        is_owner: true,
        created_at: createdAtHint || null,
      });
    }

    return rows;
  })()`, 'home-link probe');
    if (!Array.isArray(raw))
        throw new CommandExecutionError('NotebookLM home-link probe returned malformed Browser Bridge data');
    return raw.map((row) => {
        if (!isPlainObject(row) || typeof row.id !== 'string' || typeof row.title !== 'string' || typeof row.url !== 'string' || (row.created_at !== null && row.created_at !== undefined && typeof row.created_at !== 'string') || (row.is_owner !== undefined && typeof row.is_owner !== 'boolean')) {
            throw new CommandExecutionError('NotebookLM home-link probe returned a malformed row');
        }
        const url = normalizeNotebooklmNotebookUrl(row.url, row.id);
        if (!url) {
            throw new CommandExecutionError('NotebookLM home-link probe returned an untrusted or mismatched notebook URL');
        }
        return {
        id: row.id,
        title: normalizeNotebooklmTitle(row.title, 'Untitled Notebook'),
        url,
        source: 'home-links',
        is_owner: row.is_owner === false ? false : true,
        created_at: normalizeNotebooklmCreatedAt(row.created_at),
        };
    });
}
export async function listNotebooklmSourcesFromPage(page) {
    const raw = unwrapNotebooklmEvaluateResult(await page.evaluate(`(() => {
    const notebookMatch = window.location.href.match(/\\/notebook\\/([^/?#]+)/);
    const notebookId = notebookMatch ? notebookMatch[1] : '';
    if (!notebookId) return [];

    const skip = new Set([

View on GitHub (pinned to 49907e53dc)

Solutions

Reload the NotebookLM home page and retry so anchors resolve to real notebook hrefs.
Update the CLI so its link extractor filters out non-notebook anchors (buttons, section links).
Verify with your browser that the listed notebooks open at notebooklm.google.com/notebook/<id> URLs.
If scraping custom markup, ensure hrefs are resolved to absolute trusted URLs before validation.

Example fix

// before: emitting every scraped link
rows.forEach(r => out.push({ id: r.id, url: r.url }));

// after: only trusted notebook URLs
for (const r of rows) {
  const url = normalizeNotebooklmNotebookUrl(r.url, r.id);
  if (url) out.push({ id: r.id, url });
}

Defensive patterns

Strategy: validation

Validate before calling

for (const r of rows) {
  if (!/^https:\/\/notebooklm\.google\.com\/notebook\//.test(r.url) || !r.url.includes(r.id)) {
    console.warn(`Skipping untrusted notebook link for id ${r.id}`);
  }
}

Type guard

const isTrustedRowUrl = (r) => typeof r?.url === 'string' && r.url.startsWith('https://notebooklm.google.com/notebook/') && normalizeNotebooklmNotebookUrl(r.url, r.id) !== null;

Try / catch

try {
  rows = await probeHomeLinks();
} catch (e) {
  if (String(e.message).includes('untrusted or mismatched notebook URL')) {
    console.error('Reload the NotebookLM home page so links resolve to real notebook URLs.');
  } else { throw e; }
}

Prevention

Only trust absolute https notebooklm.google.com/notebook/<id> URLs.
Reload the page if you recently navigated so hrefs are not mid-transition.
Filter out non-notebook anchors (buttons, headers) before validating rows.
Never persist probe URLs without normalizing them first.

When it happens

Trigger: row.url points outside notebooklm.google.com, is a relative or javascript: link, encodes a different notebook id than row.id, or is empty while the row otherwise validated.

Common situations: Anchor elements on the home page that are not actual notebook links (menu buttons, 'New notebook' button) being scraped as rows, or SPA-router hrefs captured mid-navigation.

Related errors

AI-assisted analysis of jackwener/OpenCLI@49907e53dc (2026-08-29). Data as JSON: /api/errors/b00d67c6e4687c08. Report an issue: GitHub.