pediatric-ai-scribe-v3/test/backend-hardening.test.js

const { test } = require('node:test');
const assert = require('node:assert/strict');
const fs = require('node:fs');
const path = require('node:path');

function read(relativePath) {
  return fs.readFileSync(path.join(__dirname, '..', relativePath), 'utf8');
}

test('redactor removes PHI and common secret patterns from logs', () => {
  const { redact } = require('../src/utils/redact');
  const input = 'Authorization: Bearer abc.def.ghi password=secret123 token=tok_1234567890 email test@example.com MRN 123456789';
  const output = redact(input);

  assert.doesNotMatch(output, /abc\.def\.ghi/);
  assert.doesNotMatch(output, /secret123/);
  assert.doesNotMatch(output, /tok_1234567890/);
  assert.doesNotMatch(output, /test@example\.com/);
  assert.doesNotMatch(output, /123456789/);
  assert.match(output, /\[REDACTED\]|\[JWT\]/);
});

test('URL safety helper blocks private network targets', () => {
  const { isPrivateIp } = require('../src/utils/urlSafety');

  assert.equal(isPrivateIp('127.0.0.1'), true);
  assert.equal(isPrivateIp('10.0.0.5'), true);
  assert.equal(isPrivateIp('172.16.0.1'), true);
  assert.equal(isPrivateIp('192.168.1.10'), true);
  assert.equal(isPrivateIp('169.254.169.254'), true);
  assert.equal(isPrivateIp('100.64.0.1'), true);
  assert.equal(isPrivateIp('::1'), true);
  assert.equal(isPrivateIp('fe80::1'), true);
  assert.equal(isPrivateIp('8.8.8.8'), false);
  assert.equal(isPrivateIp('2606:4700:4700::1111'), false);
});

test('Nextcloud/WebDAV routes enforce SSRF guard and redirect blocking', () => {
  const nextcloud = read('src/routes/nextcloud.js');
  const learningAI = read('src/routes/learningAI.js');

  assert.match(nextcloud, /assertSafeHttpsUrl\(cleanUrl, 'Nextcloud URL'\)/);
  assert.match(nextcloud, /assertSafeHttpsUrl\(user\.nextcloud_url, 'Nextcloud URL'\)/);
  assert.match(nextcloud, /maxRedirects: 0/);
  assert.match(nextcloud, /encodeURIComponent\(username\)/);
  assert.match(learningAI, /assertSafeHttpsUrl\(user\.nextcloud_url, 'Nextcloud URL'\)/);
  assert.match(learningAI, /maxRedirects: 0/);
  assert.match(learningAI, /encodeURIComponent\(user\.nextcloud_user\)/);
});

test('logs and audits avoid unbounded limits and PHI-prone details', () => {
  const logs = read('src/routes/logs.js');
  const encounters = read('src/routes/encounters.js');
  const documents = read('src/routes/documents.js');
  const learningAI = read('src/routes/learningAI.js');

  assert.match(logs, /function clampLimit/);
  assert.match(logs, /clampLimit\(req\.query\.limit, 50, 200\)/);
  assert.match(logs, /logger\.warn\('client_error'/);
  assert.match(logs, /redact\(trimField\(e\.stack, 1200\)\)/);
  assert.doesNotMatch(logs, /console\.error\('\[CLIENT ERROR\]'/);

  assert.doesNotMatch(encounters, /Saved encounter: ' \+ \(label/);
  assert.doesNotMatch(encounters, /Loaded encounter: ' \+ \(row\.label/);
  assert.doesNotMatch(documents, /Uploaded: ' \+ \(req\.file/);
  assert.doesNotMatch(learningAI, /Char context/);
});

// adminMiddleware only checks req.user.role; without authMiddleware having run,
// req.user is undefined. These routes were protected only because adminConfig
// happens to be mounted on /api/admin first and guards the whole path.
test('admin routers state their own authentication, not mount order', () => {
  const fs = require('node:fs');
  const path = require('node:path');
  const root = path.join(__dirname, '..');
  for (const file of ['adminMilestones.js', 'admin.js', 'adminConfig.js', 'adminDocs.js']) {
    const src = fs.readFileSync(path.join(root, 'src/routes', file), 'utf8');
    assert.match(src, /router\.use\(authMiddleware\);/, file + ' authenticates every route it serves');
  }
  const auth = fs.readFileSync(path.join(root, 'src/middleware/auth.js'), 'utf8');
  assert.match(auth, /async function adminMiddleware\(req, res, next\) \{\s*\n\s*if \(!req\.user \|\| req\.user\.role !== 'admin'\)/,
    'and the role check stays a role check, so it fails closed on its own');
});

// Three S3 schemes grew separately — S3_*, GENERATED_IMAGES_S3_*, and later
// audio backups — which is why pointing the app at a different MinIO meant
// hunting through three files. One resolver answers it for every purpose.
test('object storage resolves the same way for every purpose', () => {
  const storage = require('../src/utils/objectStorage');

  // One endpoint plus a bucket name per purpose is enough for all of them.
  const shared = { S3_ENDPOINT: 'https://s3.example.com', S3_ACCESS_KEY: 'AK', S3_SECRET_KEY: 'SK',
    S3_BUCKET_AUDIO_BACKUPS: 'audio', S3_BUCKET_GENERATED_IMAGES: 'images', S3_BUCKET: 'docs' };
  for (const [purpose, bucket] of [['audio-backups', 'audio'], ['generated-images', 'images'], ['documents', 'docs']]) {
    const resolved = storage.settingsFor(purpose, shared);
    assert.equal(resolved.bucket, bucket, purpose + ' finds its bucket');
    assert.equal(resolved.endpoint, 'https://s3.example.com');
    assert.deepEqual(resolved.credentials, { accessKeyId: 'AK', secretAccessKey: 'SK' });
  }

  // A purpose that needs its own account still overrides everything.
  const overridden = storage.settingsFor('audio-backups',
    Object.assign({}, shared, { AUDIO_BACKUPS_S3_ENDPOINT: 'http://assets:9000', AUDIO_BACKUPS_S3_BUCKET: 'audio-backups' }));
  assert.equal(overridden.endpoint, 'http://assets:9000');
  assert.equal(overridden.bucket, 'audio-backups');

  // Existing deployments keep working untouched, including the old key names.
  const legacy = storage.settingsFor('documents',
    { S3_BUCKET: 'd', S3_REGION: 'us-west-004', S3_ENDPOINT: 'https://b2', S3_ACCESS_KEY_ID: 'A', S3_SECRET_ACCESS_KEY: 'B' });
  assert.equal(legacy.region, 'us-west-004');
  assert.deepEqual(legacy.credentials, { accessKeyId: 'A', secretAccessKey: 'B' });
  // Documents defaulted path-style off; a Backblaze endpoint must keep working.
  assert.equal(legacy.forcePathStyle, false);
  assert.equal(storage.settingsFor('audio-backups', { AUDIO_BACKUPS_S3_BUCKET: 'a', AUDIO_BACKUPS_S3_ENDPOINT: 'http://assets:9000' }).forcePathStyle, true,
    'but MinIO needs it, so a custom endpoint turns it on where there is no older default');

  // No bucket means "not configured" — never an error, since all of this is optional.
  assert.equal(storage.settingsFor('audio-backups', {}), null);
  assert.equal(storage.isConfigured('audio-backups', {}), false);

  // A mounted secret must not be overridden by an inherited environment value.
  const fs = require('node:fs'); const os = require('node:os'); const path = require('node:path');
  const keyFile = path.join(fs.mkdtempSync(path.join(os.tmpdir(), 'objstore-')), 'key');
  fs.writeFileSync(keyFile, 'from-file\n');
  const fileWins = storage.settingsFor('audio-backups',
    { AUDIO_BACKUPS_S3_BUCKET: 'a', AUDIO_BACKUPS_S3_ACCESS_KEY: 'inline', AUDIO_BACKUPS_S3_ACCESS_KEY_FILE: keyFile, AUDIO_BACKUPS_S3_SECRET_KEY: 'S' });
  assert.equal(fileWins.credentials.accessKeyId, 'from-file');
});

// .env.example listed 18 of the 67 variables the app reads, so anyone setting
// up a deployment had to find the rest by reading source.
test('.env.example documents every variable the app reads', () => {
  const fs = require('node:fs');
  const path = require('node:path');
  const root = path.join(__dirname, '..');
  const read = f => fs.readFileSync(path.join(root, f), 'utf8');

  const sources = ['server.js'].concat(
    fs.readdirSync(path.join(root, 'src'), { recursive: true })
      .filter(f => String(f).endsWith('.js')).map(f => path.join('src', String(f))));
  const used = new Set();
  for (const file of sources) {
    for (const m of read(file).matchAll(/process\.env\.([A-Z_0-9]+)/g)) used.add(m[1]);
  }
  // Set by node:test in its own child processes, never by a deployment.
  used.delete('NODE_TEST_CONTEXT');

  const documented = new Set(
    [...read('.env.example').matchAll(/^#?\s*([A-Z_0-9]+)=/gm)].map(m => m[1]));
  const missing = [...used].filter(name => !documented.has(name)).sort();
  assert.deepEqual(missing, [], 'undocumented variables: ' + missing.join(', '));
});

// registration_enabled is open-or-closed. Invite-only is the middle setting,
// and it has to hold up against someone probing codes.
test('registration invites are single-use, expiring, and safe to store', () => {
  const fs = require('node:fs');
  const path = require('node:path');
  const root = path.join(__dirname, '..');
  const read = f => fs.readFileSync(path.join(root, f), 'utf8');
  const invites = require('../src/utils/registrationInvites');
  const migration = read('migrations/1780100000000_registration-invites.js');
  const auth = read('src/routes/auth.js');

  // A code must not be recoverable from a database dump.
  assert.match(migration, /code_hash TEXT NOT NULL UNIQUE/);
  assert.doesNotMatch(migration, /\bcode TEXT\b/, 'the code itself is never stored');
  assert.equal(invites.hash('abcd-efgh'), invites.hash('ABCDEFGH'), 'hashing normalises case and separators');
  assert.notEqual(invites.hash('a'), 'a');

  // Generated codes avoid characters that are misread when typed from a screen.
  const code = invites.generateCode();
  assert.match(code, /^[0-9A-Z]{4}-[0-9A-Z]{4}-[0-9A-Z]{4}-[0-9A-Z]{4}$/);
  assert.doesNotMatch(code, /[ILOU]/, 'no I, L, O or U');
  assert.notEqual(invites.generateCode(), invites.generateCode());

  // Expiry is bounded, and a nonsense value falls back rather than throwing.
  assert.equal(invites.ttlDays(undefined), invites.DEFAULT_TTL_DAYS);
  assert.equal(invites.ttlDays(0), invites.DEFAULT_TTL_DAYS);
  assert.equal(invites.ttlDays('nonsense'), invites.DEFAULT_TTL_DAYS);
  assert.equal(invites.ttlDays(10000), invites.MAX_TTL_DAYS);

  // The claim is one conditional UPDATE, so two registrations racing the same
  // code cannot both succeed.
  const claim = read('src/utils/registrationInvites.js');
  assert.match(claim, /UPDATE registration_invites SET used_at = NOW\(\), used_by = \$1 /);
  assert.match(claim, /WHERE code_hash = \$2 AND used_at IS NULL AND revoked_at IS NULL AND expires_at > NOW\(\)/);
  assert.match(claim, /RETURNING id/);

  // Losing that race must not leave a free account behind.
  assert.match(auth, /if \(!claimedInvite\) \{[\s\S]{0,120}DELETE FROM users WHERE id = \?/);
  // And the rejection must not say which of the four reasons applied.
  assert.match(auth, /may have expired, been revoked, or already been used/);
});

// With several admins, everything in the panel was editable by all of them.
// Lockdown separates running the service from changing how it behaves.
test('admin lockdown refuses configuration writes at the server', () => {
  const lockdown = require('../src/utils/adminLockdown');
  const off = {};
  const on = { ADMIN_LOCKDOWN: 'true' };

  // It is an environment variable, not a setting: a setting could be switched
  // off by the very admin it restrains.
  assert.equal(lockdown.enabled(off), false);
  assert.equal(lockdown.enabled({ ADMIN_LOCKDOWN: 'false' }), false);
  assert.equal(lockdown.enabled(on), true);

  // Off, nothing is locked.
  assert.equal(lockdown.isLocked('prompt.hpi', off), false);

  // On, configuration is locked and day-to-day operation is not.
  for (const key of ['prompt.hpi', 'clinical_assistant.chat_model', 'models.default',
    'tts.voice', 'stt.model', 'embeddings.model', 'smtp.host', 'email.verify.subject']) {
    assert.equal(lockdown.isLocked(key, on), true, key + ' is locked');
  }
  for (const key of ['announcement.text', 'registration_enabled',
    'registration_invite_only', 'feature.memories', 'site.name']) {
    assert.equal(lockdown.isLocked(key, on), false, key + ' stays editable');
  }

  // A setting added later is locked until someone decides otherwise, rather
  // than defaulting to open.
  assert.equal(lockdown.isLocked('something.invented.later', on), true);

  const fs = require('node:fs');
  const path = require('node:path');
  const admin = fs.readFileSync(path.join(__dirname, '..', 'src/routes/adminConfig.js'), 'utf8');
  // One gate, not a check per route: a route added later cannot miss it.
  assert.match(admin, /if \(!lockdown\.enabled\(\) \|\| req\.method === 'GET' \|\| req\.method === 'HEAD'\) return next\(\);/);
  assert.match(admin, /if \(OPERATIONAL_WRITE\.test\(req\.path\)\) return next\(\);/);
  assert.match(admin, /return res\.status\(403\)\.json\(\{ error: lockdown\.refusal/);
  // And the per-key rule for the generic settings route.
  assert.match(admin, /if \(lockdown\.isLocked\(key\)\) \{\s*\n\s*return res\.status\(403\)/);

  // The panel learns the state from a response it already fetches, rather than
  // adding a request of its own on every admin open.
  assert.match(admin, /invites: await invites\.list\(\), inviteOnly: await invites\.inviteOnly\(\), lockdown: lockdown\.state\(\)/);
  const panel = fs.readFileSync(path.join(__dirname, '..', 'public/js/admin.js'), 'utf8');
  assert.match(panel, /window\.applyAdminLockdown\(data\.lockdown\)/);
});

// A citation that resolves to nothing is never rendered as a link, so without
// counting it, nobody ever learns it happened.
test('citation quality is measured on the server, where answer and sources both exist', () => {
  const audit = require('../src/utils/citationAudit');

  const sources = [{ number: 1 }, { number: 2 }, { number: 3 }];
  const good = audit.audit('Bronchiolitis is managed supportively [1]. Steroids do not help [2, 3].', sources);
  assert.equal(good.cited, 3);
  assert.deepEqual(good.unverifiableNumbers, []);

  // The case that motivated this: a number no source matched.
  const over = audit.audit('Supportive care [1] and also this [7].', sources);
  assert.equal(over.cited, 2);
  assert.deepEqual(over.unverifiableNumbers, [7]);

  // Repeats are one problem, not three.
  assert.deepEqual(audit.audit('[7] and [7] and [7]', sources).unverifiableNumbers, [7]);

  // A bracketed number inside code is content, not a citation.
  assert.deepEqual(audit.audit('```\nconst a = arr[7];\n```', sources).unverifiableNumbers, []);
  assert.deepEqual(audit.audit('Use `arr[7]` here.', sources).unverifiableNumbers, []);

  // Sources are matched on their assigned number, the same way rendering does.
  assert.deepEqual(audit.audit('[5]', [{ number: 5 }]).unverifiableNumbers, []);

  // Recording must never be able to fail an answer.
  assert.doesNotThrow(() => audit.record(null, null));
  assert.doesNotThrow(() => audit.record('[1]', undefined));

  const fs = require('node:fs');
  const path = require('node:path');
  const route = fs.readFileSync(path.join(__dirname, '..', 'src/routes/clinicalAssistant.js'), 'utf8');
  // Not awaited: the clinician is waiting for the answer.
  // One tracker for both answer paths — the streaming one and the fallback the
  // client uses when streaming fails. Auditing only the first would make
  // exactly the answers produced under failure invisible.
  assert.match(route, /function trackCitations\(req, question, answer, sources\)/);
  assert.match(route, /trackCitations\(req, prepared\.message, answer, safeSources\);/, 'streaming');
  assert.match(route, /trackCitations\(req, prepared\.message, answer, fallbackSources\);/, 'fallback');
  // Loaded on demand and allowed to be absent: observation must never be able
  // to fail an answer.
  assert.match(route, /function citationTracker\(\)/);
  assert.match(route, /if \(!tracker\) return;/);
  assert.match(route, /tracker\.store\(req\.user\.id, question, result, sources\);/);
});

test('slides are built by pandoc, from markdown, with a reference template', () => {
  // pptxgenjs is gone. It stretched every image — reading the slide XML it
  // emitted showed the target box verbatim with <a:stretch/>, so a 200x800
  // image in an 11.8x3.9 box came out 1:4 squashed to 3:1 — and it could not
  // do better, because it never measures an image; its own getSizeFromImage is
  // commented out as unused. pandoc measures them: rendered and inspected, a
  // 300x175 source produced aspect 1.714 and a 160x360 produced 0.445.
  //
  // Removing it also took image-size with it, and with that both high-severity
  // advisories (GHSA-w3rx-r6r6-pgpr, GHSA-5p2g-fcmc-qvqq), which had no fixed
  // release to upgrade to.
  const src = read('src/routes/learningAI.js');
  assert.doesNotMatch(src, /pptxgenjs|PptxGenJS/, 'nothing builds slides by hand any more');
  assert.match(src, /execFile\('pandoc', \['deck\.md', '--reference-doc=' \+ reference, '-o', 'deck\.pptx'\]/);
  assert.match(src, /timeout: 60000/, 'a conversion cannot hang the request');

  // Design lives in the template, so restyling is editing a file in PowerPoint.
  assert.match(src, /assets', 'learning', 'slides-reference\.pptx'/);
  assert.ok(require('node:fs').existsSync(
    require('node:path').join(__dirname, '..', 'assets', 'learning', 'slides-reference.pptx')),
    'and that template ships with the app');

  // pandoc must be in the image, or every export fails at runtime.
  assert.match(read('Dockerfile'), /apk add --no-cache ffmpeg curl jq pandoc-cli/);

  const pkg = JSON.parse(read('package.json'));
  assert.ok(!pkg.dependencies['pptxgenjs'], 'and the library is no longer a dependency');
  assert.ok(!pkg.dependencies['image-size'], 'nor the parser it dragged in');
});

test('a deck can only embed images the requester owns', () => {
  // pandoc resolves an image link against the filesystem, so a markdown link
  // naming any local path would read that file into the deck. Only the images
  // fetched by id for this user are written into the working directory, and
  // every other image link is dropped rather than passed through.
  const src = read('src/routes/learningAI.js');
  assert.match(src, /var image = await require\('\.\.\/utils\/generatedImages'\)\.service\(\)\.asset\(id, req\.user\);/,
    'ownership is checked when fetching');
  assert.match(src, /return local \? '!\[' \+ alt \+ '\]\(' \+ local \+ '\)' : '';/,
    'an unresolved link is removed, not passed to pandoc');
  // The working directory is per request and always cleaned up.
  assert.match(src, /workdir = await fsp\.mkdtemp\(/);
  assert.match(src, /\} finally \{[\s\S]{0,200}rm\(workdir, \{ recursive: true, force: true \}\)/);
});