{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreigdzblwriej7wcrgdiaokp64mzgjnvea3a7nt4avxx4ymiazbviwy",
    "uri": "at://did:plc:pgryn3ephfd2xgft23qokfzt/app.bsky.feed.post/3mhqfei3btdi2"
  },
  "path": "/t/technical-compendium-malicious-pdf-taxonomy-xai-gated-defensive-pipelines/174560#post_1",
  "publishedAt": "2026-03-23T15:04:11.000Z",
  "site": "https://discuss.huggingface.co",
  "tags": [
    "https://doi.org/10.6084/m9.figshare.31827337"
  ],
  "textContent": "Hi everyone,\n\nI’m sharing my latest research on the evolution of PDF weaponization and modern detection methodologies. This compendium explores the gap between traditional malware analysis and the need for **Explainable AI (XAI)** in production security stacks.\n\n**Key research areas covered:**\n\n  * **Adversarial ML:** How structural evasion techniques (2020-2026) challenge current detection models.\n\n  * **Concept Drift:** Adaptation strategies for Windows and PDF malware with minimal samples.\n\n  * **Defense Frameworks:** Integrating static, dynamic, and symbolic execution into a multi-layered pipeline.\n\n\n\n\nI’ve included a detailed taxonomy of 8 payload families, from information stealers to APT precursors.\n\n**Full Paper (DOI):** https://doi.org/10.6084/m9.figshare.31827337\n\n**Topics:** #cybersecurity #malware-detection #XAI #explainable-ai #APT #machine-learning #infosec #PDF-security",
  "title": "Technical Compendium: Malicious PDF Taxonomy & XAI-Gated Defensive Pipelines"
}