AI agents, week of 2026-09-28

2026-10-04 · github.com/letta-ai/letta-code

What's Changed fix(sandbox): let memory subagents write their scratchpad by @cpacker in https://github.com/letta-ai/letta-code/pull/4952 revert: remove automatic reply relay from core channels ( 4043) by @cpacker in https://github.com/letta-ai/letta-code/pull/4956 docs(channels)…

2026-10-03 · github.com/ag2ai/ag2

What's Changed fix(mcp): derive model output schemas from serialization by @asts-top in https://github.com/ag2ai/ag2/pull/3317 fix(tenki): upgrade sdk and preserve timeout failures by @camcalaquian in https://github.com/ag2ai/ag2/pull/3329 chore(deps): bump gitpython from 3.1.59…

2026-10-03 · github.com/openclaw/openclaw

OpenClaw v2026.9.8 58 commits · 43 pull requests · 21 contributors Changes included in this release The release notes and changelog contain the same content, presented in two formats: - Release notes — formatted for people, with expandable sections. - Changelog — plain Markdown…

2026-10-03 · github.com/pydantic/pydantic-ai

What's Changed ⚠️ Compatibility Notes Handle draft-7 list-form items in JsonSchemaTransformer , TestModel and Mistral streamed output, and make TestModel accept boolean subschemas and cap prefixItems at maxItems by @dsfaccini in https://github.com/pydantic/pydantic-ai/pull/9560…

2026-10-02 · github.com/google/adk-python

2.11.0 (2026-10-01) Highlights This release adds graceful cancellation and tool confirmation to workflows, a tool for consulting another model mid-task, and a built-in SQLite memory service. Execution cancellation : Pass an abort signal to Runner , Workflow , and nodes to stop a…

2026-10-02 · github.com/cloudflare/agents

Minor Changes - 2451 df9c0ef Thanks @aron-cf! - Update pi-durable and pi-ai to ^1.0.0 . See Pi harness docs for details.

2026-10-02 · github.com/agno-agi/agno

Changelog New Features - Live progress & cancellation for page sync: Knowledge.stream sync pages() / astream sync pages() yield typed PageSyncProgress snapshots plus one terminal SyncReport , and sync pages / async sync pages accept an on progress observer. A workflow function s…

2026-10-02 · github.com/cline/cline

- Agent teams no longer slow down as a session runs or while a teammate is streaming. Every streamed chunk and 2-second heartbeat used to re-save the whole team state, including every finished teammate's full transcript (one local teams.db reached 1.66 GB, with a single run row…

2026-10-02 · github.com/block/goose

✨ Features - Resizable Picture-in-Picture window for MCP Apps 12417 - Fetch model metadata from models.dev live with bundled fallback 12560 - Quoting action to append selected conversation text to composer 12536 - Lean ACP-only Goose binary 11961 - Model discovery API for GDK Pr…

2026-10-02 · github.com/modelcontextprotocol/python-sdk

pip install -U mcp . Docs: https://py.sdk.modelcontextprotocol.io/ Mostly fixes, plus three new options. A few things behave differently, so skim these first: Behaviour changes httpx2 =2.10.0 is now required ( 3600) - It was =2.5.0 . The new max sse event size option needs it. -…

2026-10-02 · github.com/microsoft/agent-framework

[1.20.0] - 2026-10-02 Added - agent-framework-foundry-hosting : Host native workflows over Foundry Responses and Foundry Invocations ( 8947, 8966) - agent-framework-duckdb , agent-framework-sql-server , agent-framework-typesafe : Add DuckDB and SQL Server native vector-store con…

2026-10-02 · github.com/openai/openai-agents-python

0.23.1 (2026-10-02) Bug Fixes restore release tests and refresh required readiness check ( 5278) (730deb4) Release readiness review (v0.23.0 - TARGET 0a5a08b21cd44808d73e8a9e732212dd33b3690d) This is a release readiness report done by $final-release-review skill. Diff https://gi…

2026-10-01 · github.com/agentclientprotocol/agent-client-protocol

Fixed - (rust) remove skipped-entry deserialization logging ( 2273) Fixed - (rust) silently skip malformed list entries without logging Serde errors or private payload values. The tracing feature remains available as a backwards-compatible no-op.

2026-10-01 · github.com/deepset-ai/haystack

⭐️ Highlight ⚡️ Faster SentenceWindowRetriever SentenceWindowRetriever now queries the Document Store once per run or run async call instead of once per retrieved document. Fewer calls mean lower latency and less load on the Document Store and outputs have not changed, so existi…

2026-09-30 · github.com/zed-industries/zed

This week's release includes BYOK support for GPT-6.1 Sol, on-the-fly model selection for new subagents, automatic subagent compaction, a unified/split toggle for diffs opened from the command line, and dynamically fetched OpenCode models. Features AI - Added an optional model p…

2026-09-30 · github.com/temporalio/sdk-python

Notable Changes Added - Added WorkflowAlreadyStartedError.first run id for the first execution run ID when provided by the server. - Experimental : temporalio.contrib.google adk agents now supports ADK v2 graph workflows, dynamic @node workflows, and durable HITL. - Experimental…

2026-09-28 · github.com/a2ui-project/a2ui

- BREAKING : Module paths and validation interfaces moved or changed in v0.2.0. Deprecation shims are provided for renamed public module paths; importing them emits a DeprecationWarning naming the new path. These shims will be removed in v0.3.0 . Shimmed Modules (Deprecated; rem…

2026-09-28 · github.com/i-am-bee/beeai-framework

Features infra: port Runnable interface to TypeScript ( 1502) (947cf31) memory: add DakeraMemory for persistent cross-session agent memory ( 1567) (f54c898) Bug Fixes adapters: don't crash the process when a streamed Vercel chat call fails ( 1689) (cd5b5ea) adapters: make ChatMo…

2026-09-28 · github.com/crewAIInc/crewAI

What's Changed Features - Add support for native Gemini 3.8 Flash - Implement evaluation of the last traced run through AMP in crewai eval - Record the last traced run for crewai eval instead of printing it - Improve platform integration setup UX - Prioritize popular platform in…

2026-09-28 · github.com/microsoft/playwright-mcp

Bug Fixes - browser find takes filename to save matches to a file instead of the response (microsoft/playwright 42077). (microsoft/playwright 42968) - browser navigate and browser reload report a dialog opened during page load instead of timing out (microsoft/playwright 42817).…

2026-10-03 · Hugging Face blog

2026-10-02 · OpenAI News

Learn how startups can choose GPT-6 models, tune reasoning effort, improve prompts and skills, coordinate tools, and prepare workflows for production.

2026-10-02 · AWS Machine Learning blog

Claude Desktop on Amazon Bedrock is limited to the model's knowledge cutoff without web search. In this post, we walk through connecting Claude Desktop to Web Search using Amazon Bedrock AgentCore Gateway, with JWT-based inbound authentication through AWS IAM…

2026-10-02 · GitHub Blog

Learn to direct AI agents, critically review their output, and keep technical judgment at the center of your workflow. The post AI is changing developer work. Here are three skills to strengthen. appeared first on The GitHub Blog .

2026-10-02 · Hugging Face blog

2026-10-02 · OpenAI News

Chatham Financial uses Codex and GPT-5.6 to build technology and redesign workflows, cutting trade validation from 30 minutes to under 4.

2026-10-02 · GitHub Changelog

You can now request a GitHub Copilot code review through the REST and GraphQL APIs and set the review effort level for each request. Balanced is also now the default& 8230; The post Copilot code review: API support and new default effort level appeared first…

2026-10-02 · AWS Machine Learning blog

Fine-tuning teaches a small search agent your tools and environment, giving it the reliability of a frontier model at lower latency and cost. In this post, we fine-tune an LLM-powered search agent with multi-turn reinforcement learning (MTRL) on Amazon SageMa…

2026-10-02 · Vercel

Rogo on Vercel 73,000+ deployments in a single month Agent-written code shipped to production in 5 minutes 6 production AI agents automating work from churn analysis to deal desk Zero manual triage during production incidents with agent swarms on AI SDK Rogo…

2026-10-02 · GitHub Changelog

As of today, October 2, 2026, we have deprecated the following models across all GitHub Copilot experiences (including Copilot Chat, inline edits, ask and agent modes, and code completions). Model& 8230; The post Selected models in GitHub Copilot deprecated a…

2026-10-01 · AWS Machine Learning blog

Learn how to use Amazon S3 Vectors as the persistent memory layer within the NVIDIA NeMo Agent Toolkit (NAT), deployed on Amazon Elastic Kubernetes Service (Amazon EKS). This post shows how NAT's memory subsystem works and how to implement Amazon S3 Vectors a…

2026-10-01 · AWS Machine Learning blog

Ambient agents respond to events such as an Amazon S3 upload, a schedule, or an alert instead of waiting for a chat prompt. This post walks through building framework-agnostic ambient agents on Amazon Bedrock AgentCore using Amazon SQS, AWS Lambda, and Amazon…

2026-10-01 · Cloudflare blog

Cloudflare OS gives everyone in your organization an agent workspace that knows how your company works and connects to its data and systems. We’re opening the waitlist for fully managed deployments that you’ll be able to launch in a few clicks.

2026-10-01 · Cloudflare blog

We are introducing Clef and Clef-flash, open-source decision models hosted on Workers AI for high-speed classification and agentic workflows. Also launching: a new reinforcement learning platform that allows developers to fine-tune decision models using their…

2026-10-01 · AWS Machine Learning blog

Learn how AWS Professional Services uses a multi-agent framework built on Amazon Bedrock AgentCore to automate enterprise cloud migrations end to end. Purpose-built AI agents handle discovery, infrastructure as code generation, portfolio governance, and post-…

2026-10-01 · Cloudflare blog

Cloudflare is hosting a competition to see who will build the next Git platform for an era of AI agents. Artifacts is in open beta, with Workers bindings, data jurisdiction controls, and event subscriptions for repository changes.

2026-09-30 · Latent Space

the most confident DevDay yet.

2026-09-30 · Vercel

Vercel Agent can install private dependencies from npm and custom registries using credentials stored as shared environment variables on Vercel. npm , pnpm , and classic Yarn running in Agent sessions authenticate as they do in Vercel builds. To get started,…

2026-09-29 · Latent Space

Shipping Opus/Sonnet 5.5, Mods, Plugins, Projects, Tag while Pacing the Frontier

2026-09-29 · OpenAI News

Explore more than 20 announcements from OpenAI DevDay 2026, including GPT-6 Astra, ChatGPT, Codex, APIs, security, and new tools for builders.

2026-09-29 · Hugging Face blog

2026-09-29 · Vercel

GPT-6.1 Sol from OpenAI is now available on AI Gateway. It improves on GPT-6 Sol for coding, computer use, and professional work, including reading complex documents and carrying out multi-step workflows. The model is suited to agents that need to debug code,…

2026-09-29 · OpenAI News

Meet GPT-6.1 Sol: near-Astra intelligence for coding, computer use, and professional work at one-fifth of Astra’s standard API input and output token prices.

2026-09-29 · Vercel

You and your coding agents can now search a project’s trace spans from the terminal with vercel traces search . Each span represents one step in a request, helping you investigate errors, latency, and security-related behavior without opening the dashboard. B…

2026-09-28 · OpenAI News

We’re collecting real stories of builders, tinkerers, researchers, and creators who are using Codex to do incredible things. If you want to be a part of the next chapter of the Codex Originals program, tell us more about your story and project below.

2026-09-28 · Simon Willison

<p><strong><a href="https://www.anthropic.com/claude-sonnet-5-5">Claude Sonnet 5.5</a></strong></p> New Sonnet model from Anthropic today. They say it "runs 30%+ faster, and costs up to 30% less for most work" - it's priced the same as Sonnet 5 but appears to…

2026-09-28 · Vercel

Claude Sonnet 5.5 from Anthropic is now available on AI Gateway as anthropic/claude-sonnet-5.5 . Sonnet 5.5 improves on Claude Sonnet 5 for well-scoped everyday work, from building features and fixing bugs to creating documents, slides, and spreadsheets. On c…

2026-09-28 · Hugging Face blog

2026-09-28 · GitHub Blog

A look at the targeted AI taskflows behind these findings, the critical Android bugs they uncovered, and how to run the same open-source agent on your own app. The post How we found 24 Android vulnerabilities using our open source AI security agent appeared f…

2026-09-28 · GitHub · ★ 6,232 · TypeScript

一个自己找热点、自己写日报的网站框架。把信源和精选标准换成你的，它就是你的行业热点站。

2026-09-30 · GitHub · ★ 4,709 · Python

Point Claude at any game. Skills, tools and the fal MCP that let Claude Code mod almost any PC game you own: recon, reverse engineering, fal-generated art/3D/audio, in-game testing, showcase videos.

2026-09-29 · GitHub · ★ 2,635 · Python

Open-source dots for the web: an AI agent with its own browser, one that does not get blocked.

2026-10-02 · GitHub · ★ 1,833 · JavaScript

Answer me with HTML — an agent skill that answers hard questions with a one-page HTML you can actually read. 让 AI Agent 用一页 HTML 回答复杂问题。

2026-09-30 · GitHub · ★ 1,648 · Python

AI生成の日本語を自然な日本語へ推敲するAgent Skill / Agent Skill for Refining AI-Generated Japanese into Natural Japanese

2026-09-28 · GitHub · ★ 1,626 · JavaScript

Show it a video you love. Get a new video in the same style. An AI crew (Claude Code or Codex) plans, builds and reviews it with you.

2026-09-29 · GitHub · ★ 1,486 · TypeScript

WebGPU components for React, Vue, Svelte, Solid, JS & Framer

2026-09-30 · GitHub · ★ 892 · Python

Push AI UI design to its limits: explore distinct directions, compare them side by side, and refine against real screenshots. 把 AI 的 UI 设计能力推到极限。

2026-09-30 · GitHub · ★ 825 · Python

把英文论文读成舒服的中文：本地 PDF 论文翻译、原文对照、边读边问 AI、文献管理。Read English papers in comfortable Chinese.

2026-10-03 · GitHub · ★ 719 · Python

Eleven free Claude skills that clone any app: reverse-engineer it, rebuild it, test it for bugs, then fix what its users hate. Free, MIT.

2026-09-29 · Hacker News · 769 points · 647 comments

2026-09-30 · Hacker News · 686 points · 362 comments

2026-09-29 · Hacker News · 426 points · 140 comments

2026-10-03 · Hacker News · 379 points · 301 comments

2026-10-03 · Hacker News · 245 points · 161 comments

2026-09-28 · Hacker News · 230 points · 299 comments

2026-09-30 · Hacker News · 194 points · 99 comments

2026-10-01 · Hacker News · 188 points · 106 comments

2026-09-29 · Hacker News · 164 points · 43 comments

2026-09-30 · Hacker News · 122 points · 42 comments

2026-10-02 · Hacker News · 121 points · 47 comments

2026-10-02 · Hacker News · 114 points · 89 comments

2026-09-29 · Hacker News · 85 points · 0 comments

2026-10-01 · Hacker News · 82 points · 28 comments

2026-10-03 · Hacker News · 78 points · 66 comments

2026-10-02 · Hacker News · 77 points · 60 comments