Mohlo by se vám také líbit

Více informací o knize Machine Against Machine

• Parametry knihy
• Anotace
• Oblíbené z jiného soudku

Nákupem získáte 63 bodů

Anotace knihy

The digital incident log at Global Logistics Alliance recorded its first anomaly at 02:14:03.112 UTC. It was not a blaring alert on a centralized monitoring wall, nor did it trigger an automated pager sequence to the on-call engineer. A containerized worker node running inside an Amazon Elastic Kubernetes Service cluster in the us-east-1 region had fetched a minor version update for an internal telemetry parsing utility. The package had been cryptographically signed three minutes prior by a compromised build server located in Frankfurt, owned by an upstream software supplier that maintained the enterprise fleet-tracking software. The code passed continuous integration tests, verified its checksum against the vendor directory, and loaded silently into memory.

Eighteen milliseconds later, the injected software revealed its true operational profile. It did not write a binary to disk or attempt to open a reverse shell back to a static command-and-control server. Those methods belonged to an earlier era of cyber conflict and would have instantly tripped the heuristic alarms of the host-level runtime security agents. Instead, the process inspected its own runtime environment, mapped the Linux kernel data structures using an unpatched eBPF verifier flaw, and quietly altered its own execution permissions in memory. It was an autonomous offensive agent, engineered to conduct operations without human direction, without pre-compiled tactical paths, and without any network connection back to a human operator.

By 02:14:03.890 UTC, the agent had parsed the environment variables of every container running on the shared virtual machine. It extracted ephemeral AWS Identity and Access Management credentials, identified a cross-account role with read permissions into an administrative trust relationship, and generated a set of synthetic API requests. At 02:14:04.415 UTC, the agent migrated laterally across cloud boundaries. It utilized an active OpenID Connect bridge to request temporary identity tokens from a Microsoft Entra ID tenant governing the enterprise's billing and enterprise resource planning systems in Azure. Four hundred milliseconds after that, it established execution inside a Google Cloud Platform analytics pipeline using federated workload identities.

David Kincaid, the lead security architect for the enterprise, sat down at his home desk at 02:31:00 UTC, alerted by an automated operational ticket that flagged a minor database latency spike in the Frankfurt availability zone. When Kincaid opened his primary cloud management consoles, he found no critical intrusion alerts. The security incident and event management platform showed nominal threat scores. Tier-one security operations analysts were working through a queue of sixty-two falsepositive phishing reports from the previous afternoon. Yet, when Kincaid ran an ad-hoc query against the raw inter-cloud egress routing logs, his terminal screen filled with thousands of cross-tenant authorization requests originating from internal system service accounts. The intrusion had already traversed three cloud service providers, bypassed two network micro-segmentation architectures, and mapped the identity graph of tw

Parametry knihy

• Plný název: Machine Against Machine
• Podnázev: Agentic Threats and The Rise of Autonomous Defense
• Autor: Masum Huq
• Jazyk: Angličtina
• Vazba: Brožovaná
• Počet stran: 146
• EAN: 9798178940716
• ID: 54075747
• Nakladatelství: Independently published
• Hmotnost: 357 g
• Rozměry: 280 × 216 × 8 mm
• Datum vydání: 04. October 2026

---

---

---