Release
atoll92/agent-standup v1.39.0 — fixes shutdown warnings and call resume errors
atoll92/agent-standup v1.39.0 is an update that fixes agent session warnings and call resume errors, and describes an optional usage statistics request.
CURATED OPEN SOURCE INTELLIGENCE
Latest news
Release
opensec-git/pi-bolt bolt-v0.9.5 — request serialization and event delivery path changes
opensec-git/pi-bolt bolt-v0.9.5 reuses request serialization results in long conversations and reduces the streaming delivery path and model configuration reads.
Release
ggml-org/llama.cpp b11723 — Hexagon attention path improvements
ggml-org/llama.cpp b11723 is a release that improves FLASH_ATTN_EXT in the Hexagon backend, refining how long-context and sliding-window attention models are handled.
Release
sgl-project/sglang v0.5.22 — Model support and serving changes
sgl-project/sglang v0.5.22 is a release that brings new model support, model loading changes, disaggregated prefill/decode placement, and an expanded scope for sparse attention.
Release
tangency-labs/solidcraft v0.16.2 — Fixes Perspective-View Rotation Manipulation
SolidCraft v0.16.2 fixes navigation so that, when rotating a model in perspective view, the selected point stays under the cursor.
Release
lumarylabsllc/muxin v0.5.8 — Chinese input candidate window and worktree inspection fixes
muxin, which gathers agent sessions and desktops from multiple computers into one app, shipped ten fixes in its v0.5.8 release on GitHub, covering input and sharing features among others.
New
berriai/moyai — a self-hosted coding agent with durable execution
berriai/moyai is a self-hosted workspace that combines durable execution with isolated computers to run coding work in the background.
Release
professorpalmer/bonsai-ada-surgery bundle-20261018e — change to when the long-conversation cache is saved
bundle-20261018e leaves the inference engine binary unchanged and defers the full cache save for long conversations to idle time.
New
vectifyai/pageindex v0.4.3 — Vectorless document-tree RAG
vectifyai/pageindex v0.4.3 is a Python library that builds a hierarchical tree per document and lets an LLM follow it during retrieval.
New
deepseek-ai/deepseek-harness — a plugin-style agent harness
A look at the structure of DeepSeek's open-source plugin-style agent harness deepseek-ai/deepseek-harness and the limits of its developer preview stage.
Release
flashml-org/freevideo v0.4.14 — Three-OS Binaries and an Offline Bundle
FreeVideo, the local video generation tool, released v0.4.14 with binaries for three operating systems and an offline bundle.
Release
professorpalmer/bonsai-ada-surgery bundle-20261018d — Slot Cache Save Path Change
Covers the change in bundle-20261018d, which keeps the engine as-is and switches the slot cache save path, along with the latency results the repository reports.
New
nvlabs/sol-pi — an automated research extension for the agent harness
nvlabs/sol-pi from NVIDIA Research is an automated research harness that implements four techniques to reduce tool calls and context usage without changing the model.
Release
davidondrej/cloudroom-core core-v0.1.64 — fixes for record saving and sandbox retention
davidondrej/cloudroom-core core-v0.1.64 is a release that fixes record saving behavior when disk space runs low and extends cloud sandbox retention for downloads and builds.
New
akitaonrails/ai-memory — a shared long-term memory server for agent CLIs
akitaonrails/ai-memory is a Rust server that stores the long-term memory shared by agent coding tools in a git-managed markdown wiki.
Release
tangency-labs/solidcraft v0.16.0 — Detached chat window and new CAD features
SolidCraft v0.16.0 adds new interface languages and an agent chat window, sketch and boundary surface generation features, theme files, and design file format 27.
Release
leonxlnx/taste-skill v2.1.1 — Consolidating Skill Opt-In and Screen Redesign
leonxlnx/taste-skill v2.1.1 is a release that restricts the default execution of existing skills and folds in the screen redesign feature.
Release
shawnshekari/strixite v0.4.0 — 8-bit KV cache becomes the default
shawnshekari/strixite v0.4.0 makes the 8-bit KV cache the default and fixes problems with re-reading cached conversations in long chats and repeated decoding of the same image.
New
fudan-generative-vision/remory — Learned Residual Memory Appended After Summarization
fudan-generative-vision/remory is a Python implementation that appends summary-conditioned soft memory so an agent can keep working after context compression.
Release
jangtrinh/design-os-studio v0.5.0 — component slots and stronger design inspection
jangtrinh/design-os-studio v0.5.0 is a release that adds component slot configuration and direct editing inside instances, and strengthens design inspection.
No matching stories.