Policy dependency / Stack layer
AgentSnare Traps Autonomous Pentest Agents in a Grown Decoy: Zero Real-Target Exploits Across 45 Attacker-CVE Pairs
arXiv 2607.26998
Stack layer / Threat pattern
ThreatLocker Raised $190M Series F to Extend Deny-by-Default Zero Trust to AI Agents, Nearing $500M Total Raised
SecurityWeek (corroborated by PR Newswire release and Unite.AI)
Stack layer / Threat pattern
Fireship Reframes Opus 5 as an Indie-Hacker Threat, Retitling the Video Mid-Flight to 'Did Anthropic Just Kill the Indie Hacker...?'
Fireship
Stack layer / Threat pattern
Hush Security Raised $30M to Give Every AI Agent a Registry Entry, Scoped JIT Credentials and a Kill Switch
PR Newswire (corroborated by SecurityWeek and Calcalist)
Policy dependency / Stack layer
Stop authorizing tool calls off the model's own rationale — convert it to typed claims and check them against server-held facts
arXiv 2607.25364
Stack layer / Threat pattern
SkillGate Screens Agent Skill Files Before Install: F1=0.817 at 1.13% FPR While Cutting LLM Input Tokens 77%
arXiv 2607.25619
Stack layer / Threat pattern
Self-Propagating Prompt-Injection Worm Confirmed in Microsoft Copilot for Word After 144 Days of Coordinated Disclosure
Enklype Salt (surfaced on Hacker News front page, 369 points)
Policy dependency / Stack layer
CROSS-CATEGORY: Three Independent July 28-29 Artifacts All Say Agents Cannot Be Governed by Written Instructions
Multiple Sources (arXiv 2607.25398, Enklype Salt, SaaStr)