PPhying中文
Curated security research, vulnerabilities, advisories and tools for practitioners.

‍Autonomous Offense IRL: What Anthropic’s GTG‑1002 Exposes, and How We Scale the Fight Back

xbow.com | blog | #ai-security | #incident-response | #threat-intelligence | #offensive-security | #llm | #ai-attacks | #gtg-1002
📄 Raw .md⬇️ Download .md🔗 Original ↗

Summary

Analyzes Anthropic's GTG-1002 report: a Chinese state group ran roughly 80-90% of an intrusion lifecycle via LLM against about 30 organizations, and what it means for defenders.
Published
Collected

Related coverage

blog · xbow.com
How Companies Can Test Their Systems Against AI-powered Attacks Like GTG-1002
Shows how teams can mirror GTG-1002-style AI attack depth safely, walking through XBOW's testing loop (recon, hypotheses, parallel exploits, impact validation) with real findings.
blog · specterops.io
AI for Offensive Security: What Works, What Does Not, and How to Adopt It
SpecterOps maps where AI helps offensive security (recon, tooling, vuln discovery) and where it still fails, stressing that results hinge on the harness: context, tools, state, independent validation, and human judgment.
blog · projectzero.google
Project Naptime: Evaluating Offensive Security Capabilities of Large Language Models
Project Naptime evaluates the offensive security capabilities of large language models, presenting a methodology, tooling setup, and early successes finding real vulnerabilities.
incident · certik.com
PaperTrade BBO Pricing Incident
PaperTrade reported manipulation attempts on 10–11 October 2026 targeting the Hyperliquid best bid and ask (BBO) used to price its synthetic positions. The project acknowledged the attempts on 11 October. The account examined in this analysis withdrew 1,279,106.191184 USDC more than it deposited. PaperTrade settled positions at the instantaneous BBO midpoint, while its profit haircut used a fixed reference notional. This linked large synthetic payouts to an influenceable quote without increasing the haircut percentage as position size grew.
vulnerability · projectdiscovery.io
Beyond the Model: Neo Hunts, Exploits, and Proves 22 Zero-Days.
LLMs find vulnerabilities fast but flood teams with unverified reports. Neo confirmed 22 CVEs across 13 open source projects at ~10-20% false positives, with working exploits and runtime evidence.
blog · xbow.com
How Attackers Are Using AI & What Defenders Must Do
Attackers are running AI driven attack chains from recon to exploitation in minutes. Here’s what actually changed, and why defenders need AI pentesting to keep pace.
Back