memind/mindspace-scan.mjs

const BLOCKED_ACTIVE_PATTERNS = [
  { type: 'javascript_url', pattern: /javascript:/i },
  { type: 'inline_event_handler', pattern: /\bon\w+\s*=/i },
  { type: 'embedded_active_content', pattern: /<iframe\b|<object\b|<embed\b/i },
];

const SCRIPT_TAG_PATTERN = /<script\b([^>]*)>([\s\S]*?)<\/script>/gi;
const SCRIPT_SRC_PATTERN = /\bsrc\s*=\s*(['"])([^'"]+)\1/i;
const TRUSTED_SCRIPT_SRC_PATTERNS = [
  /^\/assets\/chart\.umd\.min\.js(?:[?#].*)?$/i,
  /^https:\/\/cdn\.jsdelivr\.net\/npm\/chart\.js(?:@[^/]+)?\/dist\/chart\.(?:umd\.)?min\.js(?:[?#].*)?$/i,
  /^https:\/\/cdnjs\.cloudflare\.com\/ajax\/libs\/Chart\.js\/[^/]+\/chart\.(?:umd\.)?min\.js(?:[?#].*)?$/i,
  /^https:\/\/unpkg\.com\/chart\.js(?:@[^/]+)?\/dist\/chart\.(?:umd\.)?min\.js(?:[?#].*)?$/i,
];

const MAX_ZIP_COMPRESSION_RATIO = 100;
const MAX_ZIP_LOCAL_HEADERS = 512;

function isZipContainer(mimeType) {
  return (
    mimeType.includes('openxmlformats') ||
    mimeType.includes('msword') ||
    mimeType.includes('ms-excel') ||
    mimeType.includes('ms-powerpoint')
  );
}

function isTrustedScriptSrc(src) {
  const normalized = String(src ?? '').trim().replace(/^\/\//, 'https://');
  return TRUSTED_SCRIPT_SRC_PATTERNS.some((pattern) => pattern.test(normalized));
}

function scanScriptTags(sample, { htmlActiveContentPolicy }) {
  const scripts = [...sample.matchAll(SCRIPT_TAG_PATTERN)];
  if (!scripts.length) return null;
  if (htmlActiveContentPolicy !== 'sandbox_warn') {
    return {
      scanStatus: 'blocked',
      riskLevel: 'high',
      findings: ['text_active_content'],
    };
  }
  const untrusted = scripts.some((match) => {
    const attrs = match[1] ?? '';
    const src = attrs.match(SCRIPT_SRC_PATTERN)?.[2] ?? null;
    return src ? !isTrustedScriptSrc(src) : false;
  });
  if (untrusted) {
    return {
      scanStatus: 'blocked',
      riskLevel: 'high',
      findings: ['text_active_content'],
    };
  }
  return {
    scanStatus: 'warned',
    riskLevel: 'medium',
    findings: ['trusted_html_active_content'],
  };
}

function scanTextContent(buffer, options = {}) {
  const sample = buffer.subarray(0, Math.min(buffer.length, 256 * 1024)).toString('utf8');
  for (const { type, pattern } of BLOCKED_ACTIVE_PATTERNS) {
    if (pattern.test(sample)) {
      return {
        scanStatus: 'blocked',
        riskLevel: 'high',
        findings: [type],
      };
    }
  }
  const scriptFinding = scanScriptTags(sample, options);
  if (scriptFinding) return scriptFinding;
  return null;
}

function scanZipStructure(buffer) {
  if (buffer.subarray(0, 2).toString('ascii') !== 'PK') {
    return {
      scanStatus: 'blocked',
      riskLevel: 'medium',
      findings: ['zip_signature_mismatch'],
    };
  }

  let offset = 0;
  let headers = 0;
  while (offset + 30 <= buffer.length && headers < MAX_ZIP_LOCAL_HEADERS) {
    if (buffer.subarray(offset, offset + 2).toString('ascii') !== 'PK') break;
    const compressionMethod = buffer.readUInt16LE(offset + 8);
    const compressedSize = buffer.readUInt32LE(offset + 18);
    const uncompressedSize = buffer.readUInt32LE(offset + 22);
    const nameLength = buffer.readUInt16LE(offset + 26);
    const extraLength = buffer.readUInt16LE(offset + 28);
    const headerSize = 30 + nameLength + extraLength;
    if (headerSize <= 0 || offset + headerSize > buffer.length) break;

    if (compressionMethod !== 0 && compressedSize > 0 && uncompressedSize / compressedSize > MAX_ZIP_COMPRESSION_RATIO) {
      return {
        scanStatus: 'blocked',
        riskLevel: 'critical',
        findings: ['zip_bomb_ratio'],
      };
    }

    headers += 1;
    offset += headerSize;
    if (compressedSize > 0) {
      offset += compressedSize;
    }
  }

  return null;
}

export function runBasicFileScan(buffer, { filename, mimeType, htmlActiveContentPolicy = 'block' }) {
  if (!Buffer.isBuffer(buffer) || buffer.length === 0) {
    return {
      scanStatus: 'blocked',
      riskLevel: 'high',
      findings: ['empty_file'],
    };
  }

  if (mimeType.startsWith('text/') || mimeType === 'application/pdf') {
    const textFinding = scanTextContent(buffer, { filename, mimeType, htmlActiveContentPolicy });
    if (textFinding) return textFinding;
  }

  if (mimeType === 'application/pdf' && buffer.subarray(0, 5).toString('ascii') !== '%PDF-') {
    return {
      scanStatus: 'blocked',
      riskLevel: 'medium',
      findings: ['pdf_signature_mismatch'],
    };
  }

  if (isZipContainer(mimeType)) {
    const zipFinding = scanZipStructure(buffer);
    if (zipFinding) return zipFinding;
  }

  return {
    scanStatus: 'passed',
    riskLevel: 'none',
    findings: [],
  };
}