Stack layer / Threat pattern
An Attacker Built a Credential-Harvesting Pipeline With a Multi-Agent Framework in Under Six Hours
Google Cloud Threat Intelligence
Policy dependency / Stack layer
Hierarchical Ransomware Agents Escalate to Dynamic and Memory Analysis Only on Specialist Disagreement
arXiv 2609.04820
Stack layer / Threat pattern
16% of 3,171 public agent-harness setups carry a confirmed security defect, and 3.8% ship a skill that pre-approves your shell
arXiv
Stack layer / Contrast
Automattic's offline grammar checker Harper carries 890 open items at 15,183 stars
GitHub Trending
Stack layer / Contrast
llama.cpp enabled speculative-decoding state rollback for Kimi-K3, which had it switched off
GitHub
Policy dependency / Stack layer
83.3% of 281 open-source AI contribution policies permit AI code, but 67.3% demand a high level of human involvement
arXiv
Stack layer
CROSS-CATEGORY: Three Independent Large-N Agent Evals Published in 48 Hours, All by Individuals, All Finding the Simple Configuration Wins
Dan Luu, Bottleneck Labs and alvins82 (three independent sources)
Policy dependency / Stack layer
MemSentry gates persistent memory writes on a signed security-state delta rather than on content classification
arXiv