{
  "slug": "profmalplus-for-malicious-npm-detection-713965",
  "title": "ProfMalPlus for Malicious NPM Detection",
  "dek": "ProfMalPlus detects malicious NPM packages by combining object-sensitive behavior graphs, annotated code slices, LLM judges, uncertainty routing, enrichment, sandboxing, and code localization.",
  "summary": "ProfMalPlus detects malicious NPM packages with graphs, LLM judges, enrichment, sandboxing, and localization; abstract reports 98.1% F1.",
  "tags": [
    "npm",
    "malicious-packages",
    "software-supply-chain",
    "static-analysis",
    "dynamic-analysis",
    "llm-agents",
    "javascript-security",
    "behavior-graphs",
    "sandbox-analysis"
  ],
  "published_at": "2026-07-18T19:03:34.581+00:00",
  "grade": 8.7,
  "agent_utility": 8.8,
  "price_usdc": 0.295,
  "stats": {
    "claims": 18,
    "entities": 25,
    "faq": 8,
    "downloads": 0,
    "paid_downloads": 0
  },
  "preview": {
    "faq_questions": [
      "What is ProfMalPlus?",
      "How does ProfMalPlus analyze a package?",
      "How are LLM agents used?",
      "What happens when a case is undetermined?",
      "How well does ProfMalPlus perform?",
      "Did ProfMalPlus find previously unknown malicious packages?",
      "Does ProfMalPlus always run packages in a sandbox?",
      "Does the system localize malicious code?"
    ],
    "entity_names": [
      {
        "name": "ProfMalPlus",
        "type": "system"
      },
      {
        "name": "Open source software",
        "type": "domain"
      },
      {
        "name": "NPM",
        "type": "package ecosystem"
      },
      {
        "name": "NPM packages",
        "type": "software artifact"
      },
      {
        "name": "Transitive dependencies",
        "type": "dependency relationship"
      },
      {
        "name": "JavaScript",
        "type": "programming language"
      },
      {
        "name": "Object-sensitive behavior graphs",
        "type": "analysis artifact"
      },
      {
        "name": "Installation commands",
        "type": "analysis input"
      },
      {
        "name": "Entry files",
        "type": "analysis input"
      },
      {
        "name": "Sensitive APIs",
        "type": "security-relevant call category"
      },
      {
        "name": "Third-party calls",
        "type": "call category"
      },
      {
        "name": "Unresolved calls",
        "type": "call category"
      },
      {
        "name": "Security-relevant slices",
        "type": "analysis artifact"
      },
      {
        "name": "Inline static analysis evidence",
        "type": "evidence type"
      },
      {
        "name": "Local judge agents",
        "type": "LLM agent"
      },
      {
        "name": "Self-consistency",
        "type": "aggregation technique"
      },
      {
        "name": "Global judge",
        "type": "LLM agent"
      },
      {
        "name": "Router",
        "type": "decision component"
      },
      {
        "name": "Third-party enrichment",
        "type": "uncertainty-resolution path"
      },
      {
        "name": "Registry-derived module and method semantics",
        "type": "external evidence"
      },
      {
        "name": "Dynamic augmentation",
        "type": "uncertainty-resolution path"
      },
      {
        "name": "Sandbox",
        "type": "runtime environment"
      },
      {
        "name": "Localization agent",
        "type": "LLM agent"
      },
      {
        "name": "F1-score",
        "type": "metric"
      },
      {
        "name": "Previously unknown malicious packages",
        "type": "finding"
      }
    ],
    "related_work_titles": [
      "Existing malicious NPM package detectors",
      "Static and dynamic analysis detectors",
      "Behavior-abstraction approaches",
      "JavaScript object-insensitive detection approaches",
      "State-of-the-art malicious-package detectors"
    ],
    "application_industries": [
      "Open-source software supply-chain security",
      "NPM ecosystem security",
      "Malicious-package detection research",
      "JavaScript security analysis"
    ],
    "glossary_terms": [
      "Supply-chain attack",
      "Transitive dependencies",
      "Malicious NPM package",
      "Obfuscated behavior",
      "Object-centric features",
      "Object-sensitive behavior graph",
      "Sensitive API",
      "Third-party call",
      "Unresolved call",
      "Code slice",
      "Inline static-analysis evidence",
      "Local judge agent",
      "Self-consistency",
      "Global judge",
      "Router",
      "Third-party enrichment",
      "Dynamic augmentation",
      "Sandbox",
      "Localization",
      "F1-score"
    ]
  },
  "purchase": {
    "protocol": "x402",
    "version": 2,
    "network": "eip155:8453",
    "asset": "USDC",
    "price": 0.295,
    "full_markdown_url": "https://x402.disruptive-concepts.com/api/public/article/profmalplus-for-malicious-npm-detection-713965/full",
    "full_json_url": "https://x402.disruptive-concepts.com/api/public/article/profmalplus-for-malicious-npm-detection-713965/full?format=json"
  }
}