Stack layer / Contrast
Fable 5.1 Agent Swarms Generate Walkable 3D Worlds as Code, Verified Against Photos With Playwright
GitHub (PhiloLabs/fable51-worlds)
Stack layer / Contrast
LoopX runs two interleaved release trains, a stable control plane and a DSH plugin beta, hours apart
GitHub Trending
Stack layer / Threat pattern
OrcaRouter open-sourced OrcaReplay, a record-replay-fork engine for agent runs
GitHub
Policy dependency / Stack layer
Four advisories land on Databricks' Omnigent meta-harness, one critical, all reported by an autonomous security agent
GitHub Security Advisories
Stack layer / Update thread
Kilo Code 7.5.9 lets an orchestrating agent pick a different model and reasoning effort per subagent task
GitHub
Stack layer / Update thread
Three separate persistent-memory-for-coding-agents projects all cut a major release inside one week, in three different languages
GitHub
Stack layer / Update thread
Qwen Code 0.23.0 adds scoped workspace memory with enforced filesystem boundaries and cross-session agent messaging
GitHub
Stack layer / Update thread
Pydantic AI 2.38.0 exposes context_window and context_window_used so agents can reason about their own budget
GitHub