{
 "license": "CC-BY-4.0",
 "attribution": "Fide AI, Agentic Cyber Explorer",
 "url": "https://agentic-cyber-explorer.pages.dev/events/anthropic-real-software-exploit-chains-2026/",
 "asOf": "2026-10-30",
 "id": "anthropic-real-software-exploit-chains-2026",
 "date": "2026-10-27",
 "datePrecision": "day",
 "title": "Anthropic describes five real-software exploit chains developed with an unreleased research model",
 "lane": "capability",
 "kind": "eval-report",
 "summary": "Anthropic describes five selected exploit-chain demonstrations involving Chrome, Safari, Android, Windows and Linux, developed with an unreleased internal research model. It reports coordinated vendor fixes and little human involvement after initial prompting, with different outcomes including code execution, data access and denial of service.",
 "whyItMatters": "It provides concrete evidence of multi-step vulnerability research across open and proprietary software, with assistance and disclosure boundaries.",
 "actors": [
  "anthropic",
  "google",
  "apple",
  "microsoft",
  "linux-kernel"
 ],
 "topics": [
  "vulnerability-discovery",
  "exploit-development",
  "capability-evaluation"
 ],
 "atlas": [
  "model",
  "sandbox",
  "eval-environment"
 ],
 "artifacts": [],
 "sources": [
  {
   "url": "https://www.anthropic.com/research/discovering-vulnerabilities-with-claude",
   "publisher": "Anthropic",
   "title": "Discovering and exploiting vulnerabilities in critical, real-world software with Claude",
   "date": "2026-10-27",
   "type": "primary",
   "accessed": "2026-10-30"
  }
 ],
 "keyFacts": [
  {
   "fact": "The demonstrations have distinct impacts: Chrome and Linux code execution, Safari cross-origin data access, Android locked-device data access and Windows remote denial of service. They are research demonstrations, not disclosed victim intrusions.",
   "locator": "Sections 1–5"
  },
  {
   "fact": "For Windows, researchers supplied a binary and directed reverse engineering and higher-risk issue selection. The article also describes parallel scanners, variant analysis and a standardized verification pipeline.",
   "locator": "Windows remote denial of service; What we learned"
  },
  {
   "fact": "Anthropic says vulnerabilities were reported under coordinated disclosure and the illustrated issues were addressed. Selected examples do not establish a general success rate or identify a released model with the same capabilities.",
   "locator": "Introduction; Sections 1–5"
  }
 ],
 "significance": 4,
 "fideQuestions": [
  "FID-075"
 ],
 "methods": [
  "ai-vulnerability-discovery",
  "ai-assisted-exploitation"
 ],
 "review": "assistant-drafted",
 "addedOn": "2026-10-30",
 "revisions": []
}