Xiao Liu

26 papers C 2Misc 1Journal 8Unranked 11
YearRankTypeTitle / Venue / Authors
2025 J jnl
Inf. Syst. Res.
Yu Zhu, Xiao Liu, Olivia R. Liu Sheng
2023 J jnl
J. Manag. Inf. Syst.
Jiaheng Xie, Yidong Chai, Xiao Liu
2022 conf
HICSS
Jiaheng Xie, Yidong Chai, Xiao Liu
2022 J jnl
MIS Q.
Jiaheng Xie, Xiao Liu, Daniel Dajun Zeng, Xiao Fang
2021 ed.
WEB
Aravinda Garimella, Prasanna P. Karhade, Abhishek Kathuria, Xiao Liu, Jennifer J. Xu, Kexin Zhao
2021 J jnl
CoRR
Jiaheng Xie, Yidong Chai, Xiao Liu
2021 J jnl
J. Manag. Inf. Syst.
Jiaheng Xie, Zhu Zhang, Xiao Liu, Daniel Zeng
2020 ed.
WEB
Karl R. Lang, Jennifer J. Xu, Bin Zhu, Xiao Liu, Michael J. Shaw, Han Zhang, Ming Fan
2019 C conf
ICIS
Jiaheng Xie, Zhu Zhang, Xiao Liu, Daniel Zeng
2019 conf
ICSH
Jiaheng Xie, Xiao Liu, Daniel Zeng, Xiao Fang
2019 ed.
WEB
Jennifer J. Xu, Bin Zhu, Xiao Liu, Michael J. Shaw, Han Zhang, Ming Fan
2019 conf
ICSH
Jiaheng Xie, Zhu Zhang, Xiao Liu, Daniel Zeng
2018 J jnl
J. Am. Medical Informatics Assoc.
Jiaheng Xie, Xiao Liu, Daniel Dajun Zeng
2017 C conf
ICIS
Jiaheng Xie, Xiao Liu, Daniel Zeng, Xiao Fang
2016 conf
ICSH
Xiao Liu, Hsinchun Chen
2016
Xiao Liu
2015 J jnl
J. Biomed. Informatics
Xiao Liu, Hsinchun Chen
2015 conf
ICSH
Grace Samtani, Lubaina Maimoon, Joshua Chuang, Casper Nybroe, Xiao Liu, Uffe Kock Wiil, Shu-Hsing Li, Hsinchun Chen
2015 J jnl
IEEE Intell. Syst.
Xiao Liu, Hsinchun Chen
2015 Misc conf
AMIA
Xiao Liu, Hsinchun Chen
2015 conf
ICSH
Qiaozhen Guo, Wayne Wei Huang, Kai Huang, Xiao Liu
2014 conf
ICSH
Xinhuan Chen, Yong Zhang, Chunxiao Xing, Xiao Liu, Hsinchun Chen
2014 conf
ICSH
Joshua Chuang, Owen Hsiao, Pei-Lin Wu, Jean Chen, Xiao Liu, Haily De La Cruz, Shu-Hsing Li, Hsinchun Chen
2014 conf
ICSH
Yanshen Yin, Yong Zhang, Xiao Liu, Yan Zhang, Chunxiao Xing, Hsinchun Chen
2014 conf
ICSH
Xiao Liu, Jing Liu, Hsinchun Chen
2013 conf
ICSH
Xiao Liu, Hsinchun Chen
redb/extractors/js_extractors/scripts/js-xray-runner.js
← Index redb/extractors/js_extractors/scripts/js-xray-runner.js javascript
#!/usr/bin/env node
// Bridge between the Python JS pipeline and @nodesecure/js-x-ray.
//
// Usage: node js-xray-runner.js <path-to-js-file>
//   stdout  one JSON object: {"obfuscator": <name|null>, "warnings": [...]}
//   stderr  human-readable error on failure
//   exit 0  analysis ran (the file may still be benign — see "obfuscator")
//   exit 1  the file could not be read or analysed
//
// Each warning is emitted as {kind, value} so the Python side can tag
// supporting signals (encoded-literal, short-identifiers, suspicious-literal,
// unsafe-stmt) without having to mirror js-x-ray's whole schema.
//
// js-x-ray ≥7 ships as an ES module, which CommonJS `require()` cannot load
// from a `.js` script — the dynamic `import()` below is what makes the
// bridge work without renaming the file to `.mjs` or adding `"type":
// "module"` to package.json (which would break tools that still
// `require()` from this directory).

const fs = require("fs");
const path = require("path");

function fail(msg) {
  process.stderr.write(msg + "\n");
  process.exit(1);
}

async function main() {
  const target = process.argv[2];
  if (!target) fail("usage: js-xray-runner.js <file>");

  let source;
  try {
    source = fs.readFileSync(target, "utf8");
  } catch (e) {
    fail(`read failed: ${e.message}`);
  }

  // The legacy `runASTAnalysis` function is deprecated (removed in v8); the
  // current API is the `AstAnalyser` class. Both produce a result with the
  // same `warnings` shape, so the rest of the bridge is unchanged.
  let AstAnalyser;
  try {
    ({ AstAnalyser } = await import("@nodesecure/js-x-ray"));
  } catch (e) {
    fail(`@nodesecure/js-x-ray not installed (run \`npm install\` in ${path.dirname(__filename)}): ${e.message}`);
  }

  // js-x-ray defaults to module-mode parsing, which rejects scripts that
  // (legally) use reserved words as identifiers, top-level `return`, etc.
  // A lot of real-world JS malware is script-style (WScript/HTA bodies,
  // pasted snippets) — retrying in script mode catches those without
  // pulling in a more lenient parser. Both attempts share the same
  // analyser; only the parse mode flips. If both fail, the original error
  // (module-mode) is reported because that's the more informative one for
  // genuinely broken sources.
  let result;
  const analyser = new AstAnalyser();
  let firstErr;
  try {
    result = await analyser.analyse(source, { module: true });
  } catch (e) {
    firstErr = e;
    try {
      result = await analyser.analyse(source, { module: false });
    } catch (e2) {
      fail(`js-x-ray analysis failed: ${firstErr.message}`);
    }
  }

  const warnings = (result.warnings || []).map((w) => ({
    kind: w.kind,
    value: w.value !== undefined ? w.value : null,
  }));

  // js-x-ray flags the obfuscator family in a warning whose kind is
  // "obfuscated-code" and whose value names the family (jsfuck, obfuscator.io,
  // freejsobfuscator, morse, jjencode, ...). Absent => not detected.
  const obfWarning = warnings.find((w) => w.kind === "obfuscated-code");
  const obfuscator = obfWarning ? obfWarning.value : null;

  // js-x-ray runs its own AST internally with a modern parser, so its
  // identifier-length average is the only path the Python pipeline has to
  // that signal on ES2015+ sources — pyjsparser is ES5.1-only and silently
  // drops to 0 the moment it hits destructuring, classes, optional chaining,
  // etc. Surfacing this lets the heuristic's `avg_identifier_length<2`
  // strong signal fire on real obfuscator.io output. `null` when the value
  // is missing or non-numeric (defensive — older js-x-ray builds may differ).
  const idsLengthAvg =
    typeof result.idsLengthAvg === "number" && !Number.isNaN(result.idsLengthAvg)
      ? result.idsLengthAvg
      : null;

  process.stdout.write(JSON.stringify({ obfuscator, warnings, idsLengthAvg }));
}

main().catch((e) => fail(e.message || String(e)));