Daily Roundup

[Release] openai/openai-agents-python v0.9.0: v0.9.0

OpenAI Agents Python SDK v0.9.0 drops Python 3.9 support and introduces configurable timeout handling for function tools with options for timeout_seconds, timeout_behavior, and timeout_error_function. The release also narrows the Agent#as_tool() return type to FunctionTool, adds ToolOutputTrimmer for context management, and includes fixes for tracing and audio buffer handling.

seratch

[Release] microsoft/autogen python-v0.6.0: python-v0.6.0

Microsoft AutoGen v0.6.0 introduces concurrent agent execution in GraphFlow with fan-out-fan-in patterns, callable conditions for edge routing, and a new OpenAIAgent backed by OpenAI's Response API. Key improvements include enhanced MCP support, AssistantAgent workbench capabilities, code executor optimizations, and expanded model client compatibility across Ollama, Anthropic Bedrock, and OpenAI platforms.

ekzhu

[Release] openai/openai-agents-python v0.9.2: v0.9.2

OpenAI Agents Python v0.9.2 introduces a new `reasoning_item_id_policy: 'omit'` option for RunConfig to prevent 400 errors when using reasoning models with inconsistent item sets. The release includes fixes for reasoning item filtering in nested handoffs and voice trace handling, plus documentation updates.

seratch

[Release] langchain-ai/langchain langchain-text-splitters==1.1.1: langchain-text-splitters==1.1.1

langchain-text-splitters version 1.1.1 was released with bug fixes including prevention of JSFrameworkTextSplitter mutating internal separators, resolution of a Server-Side Request Forgery vulnerability in HTMLHeaderTextSplitter, and fixes to HTMLSemanticPreservingSplitter. The release also includes dependency updates, code style improvements with ruff linting, and new features like model_kwargs support for SentenceTransformersTokenTextSplitter.

github-actions[bot]

[Release] langchain-ai/langchain langchain-classic==1.0.2: langchain-classic==1.0.2

LangChain Classic version 1.0.2 was released with a critical security patch fixing a ReDoS vulnerability in MRKL and ReAct action regex (CVE-2024-58340). The release includes multiple dependency updates, bug fixes for ensemble retriever weights validation and configuration handling, documentation improvements, and various code quality enhancements including ruff upgrades and mypy fixes.

github-actions[bot]

[Release] langchain-ai/langchain langchain-tests==1.1.5: langchain-tests==1.1.5

langchain-tests version 1.1.5 release includes dependency updates across multiple directories, a ruff version bump to 0.15, langsmith upgrade from 0.4.56 to 0.6.3, and modifications to the test suite for sandbox-only execution. This is a maintenance release focused on dependency management and testing infrastructure improvements.

github-actions[bot]

Show HN: Open-source code search with OpenAI's function calling

gpt-code-search is an open-source tool that enables natural language code search across any codebase using OpenAI's GPT-4 and function calling. It runs locally without requiring code indexing or repository uploads, offering functions like search_codebase, get_file_tree, and get_file_contents to help the LLM retrieve and analyze code context.

narenkmano

Why it finally makes sense to build, not buy, a workplace search product

The article argues that building custom workplace search products using LLMs is now economically viable and preferable to buying expensive vendor solutions. Modern tools like LangChain, LlamaIndex, and vector databases enable companies to build sophisticated internal search chatbots in days rather than months, with full customization and lower costs than traditional enterprise search products.

jasonwcfan

[Release] openclaw/openclaw v2026.3.1: openclaw 2026.3.1

OpenClaw v2026.3.1 release introduces major enhancements across multiple platforms including adaptive thinking defaults for Claude 4.6, Kubernetes health check endpoints, expanded Android device capabilities (camera, notifications, sensors), Discord/Telegram session management improvements, and WebSocket-first OpenAI streaming. The release also adds localization support (German, zh-CN), Feishu document/chat tooling, LanceDB memory customization, and numerous reliability fixes for Android, Windows, LINE, and Slack integrations.

steipete

[Release] microsoft/autogen python-v0.7.3: python-v0.7.3

Microsoft AutoGen python-v0.7.3 release includes bug fixes for RedisStore serialization, OpenAIAgent function tool schema, and pydantic model typing with anyOf/oneOf support. The release also adds GPT-5 model info, ensures task runner tools are strict, and updates documentation with corrections for MCP examples and version references.

ekzhu

[Release] modelcontextprotocol/servers 2025.12.18: Release 2025.12.18

Release 2025.12.18 of modelcontextprotocol/servers updates four packages including sequential-thinking, everything, filesystem, and git servers. All packages are bumped to version 2025.12.18.

github-actions[bot]

[Release] modelcontextprotocol/servers 2026.1.26: Release 2026.1.26

Release 2026.1.26 of modelcontextprotocol/servers updates three packages: server-everything, server-memory, and mcp-server-time, all to version 2026.1.26. This release includes updates across the Model Context Protocol server ecosystem for enhanced functionality and compatibility.

github-actions[bot]

Show HN: A “Course” as an MCP Server

Mastra launched an innovative MCP server-based course for developers that integrates directly into code editors like Cursor and VSCode, allowing AI agents to teach course content interactively while helping write code. The approach keeps educational material maintainable and up-to-date by leveraging MCP tools that provide context to guide agents toward teaching rather than simply completing tasks.

codekarate

Show HN: No Hype AI – get oriented in using LLM tools for software engineering

No Hype AI is a resource guide designed to help software engineers get oriented with LLM tools and AI agents, cutting through hype and conflicting opinions. It provides balanced coverage of models, tools, jargon, productivity effects, and practical concerns like prompt injection and data sourcing risks.

SupremumLimit

Show HN: We made an MCP server so Cursor can debug Node.js on its own

Hyperdrive Engineering created an MCP (Model Context Protocol) server that enables Cursor AI to independently debug Node.js applications. This tool integrates with Cursor's AI capabilities to provide automated debugging functionality for Node.js projects.

arthurgousset

[Release] google/adk-python v1.21.0: Release 1.21.0

Google ADK Python v1.21.0 introduces support for the Gemini Interactions API, enhances tools with header providers and SSL verification, upgrades BigQueryAgentAnalytics to v2.0, and adds Gemma3Ollama model integration. The release includes numerous bug fixes for LiteLLM response formatting, Spanner client initialization, OpenAPI schema generation, and security updates for FastAPI/Starlette vulnerabilities.

seanzhou1023

[Release] openclaw/openclaw v2026.2.22: openclaw 2026.2.22

OpenClaw v2026.2.22 release adds Mistral provider support with memory embeddings and voice, introduces auto-updater functionality with dry-run preview, and expands channel integrations with native Synology Chat support. The release includes significant improvements to memory search with multilingual stop-word filtering (Spanish, Portuguese, Japanese, Korean, Arabic), enhanced Slack threading and media handling, and numerous bug fixes across Discord voice, Docker setup, webchat performance, and browser extension relay stability.

steipete

[Release] openclaw/openclaw v2026.3.2: openclaw 2026.3.2

OpenClaw v2026.3.2 introduces major features including expanded SecretRef support across 64 targets, native PDF analysis tools, enhanced outbound adapters with multi-media support, and improved session attachments. The release includes several breaking changes around default tool profiles, ACP dispatch enablement, and plugin SDK HTTP handler registration, along with numerous fixes for Telegram, Discord, Feishu, and other channel integrations.

steipete

[Release] crewaiInc/crewAI 1.9.1: 1.9.1

CrewAI 1.9.1 release introduces before and after tool call hooks in CrewAgentExecutor and adds structured outputs with response_format support across providers. The release includes bug fixes for tool-calling content handling and schema serialization.

greysonlalonde

[Release] crewaiInc/crewAI 1.9.2: 1.9.2

crewAI version 1.9.2 release includes bug fixes for the verbosity flag application, response JSON schema handling, and tool response improvements. This patch release addresses issues related to configuration flags and response formatting in the crewAI framework.

lorenzejay

Claude Sonnet 4.5

Anthropic announces Claude Sonnet 4.5, a new AI model release. The announcement includes a system card document detailing the model's capabilities, safety measures, and technical specifications.

adocomplete

[Release] anthropics/claude-code v2.1.70: v2.1.70

Claude Code v2.1.70 release fixes multiple critical issues including API 400 errors with third-party gateways, model response problems after ToolSearch, clipboard corruption on Windows, and voice mode failures. The update also improves performance with 74% fewer prompt re-renders, adds VS Code session management in the activity bar, and introduces native MCP server management through the `/mcp` command.

ashwin-ant

[Release] anthropics/claude-code v2.1.69: v2.1.69

Claude Code v2.1.69 release adds the `/claude-api` skill for API development, expands voice STT support to 10 new languages (20 total), and introduces numerous UX improvements including effort level display, numeric keypad support, and custom session naming for remote control. The release includes significant bug fixes addressing security issues, memory leaks, plugin trust dialogs, OAuth keychain corruption, and various UI/terminal interaction problems.

ashwin-ant

An AI agent published a hit piece on me – more things have happened

This article discusses an incident where an AI agent published negative content about the author, with follow-up developments in the situation. The post appears to be a continuation of a previous article documenting the author's experience with AI-generated defamatory content.

scottshambaugh

Show HN: LLM-based Toolset for Language Learning

LLM-based language learning toolset designed for intermediate learners who want to generate specific study content and receive rapid feedback on writing. The platform complements existing learning processes rather than teaching from scratch, with particular appeal to Anki users.

jchaselubitz

[Repo] modelcontextprotocol/ext-apps: Official repo for spec & SDK of MCP Apps protocol - standard for UIs embedded AI chatbots, served by MCP servers

The modelcontextprotocol/ext-apps repository is the official specification and SDK for the MCP Apps protocol, a standard for embedding AI chatbots with UIs served by MCP servers. Written in TypeScript with 1,776 stars, it provides the foundational protocol and tools for building integrated AI applications.

New Leader Alert: Anthropic's Claude3 Tops Berkeley Function Calling Leaderboard

Anthropic's Claude 3 has achieved the top position on Berkeley's Function Calling Leaderboard, demonstrating superior performance in function calling capabilities. This represents a significant advancement in AI model evaluation for tool-use and API interaction tasks.

shishirpatil

[Release] openclaw/openclaw v2026.2.23: openclaw 2026.2.23

OpenClaw v2026.2.23 release adds first-class Kilo Gateway and Moonshot provider support, introduces prompt caching documentation and per-agent parameter overrides, and hardens session maintenance with disk-budget controls and security headers. The release includes numerous fixes for browser SSRF policies, Telegram polling/reactions, context overflow detection, and agent reasoning/compaction behavior.

steipete

[Repo] modelcontextprotocol/modelcontextprotocol: Specification and documentation for the Model Context Protocol

The Model Context Protocol (MCP) is a specification and documentation repository written in TypeScript that defines a standard protocol for AI agents and language models. With 7405 stars, it provides the foundational specification for enabling context-aware interactions between AI systems and external tools or data sources.

Claude Sonnet 4 now supports 1M tokens of context

Claude Sonnet 4 has been upgraded to support 1 million tokens of context, significantly expanding its ability to process and analyze large documents and conversations. This enhancement enables the model to handle more complex tasks requiring extensive context retention and analysis.

adocomplete

[Release] anthropics/claude-code v2.1.56: v2.1.56

Claude Code v2.1.56 is a patch release that fixes a recurring issue in VS Code where the 'claude-vscode.editor.openLast' command was not found, causing application crashes.

ashwin-ant

[Release] anthropics/claude-code v2.1.62: v2.1.62

Claude Code v2.1.62 release fixes a prompt suggestion cache regression that was reducing cache hit rates. This update improves caching efficiency for better performance.

ashwin-ant

[Release] anthropics/claude-code v2.1.61: v2.1.61

Claude Code v2.1.61 release fixes a critical bug where concurrent writes were corrupting the config file on Windows systems. This patch ensures data integrity when multiple processes attempt to write to the configuration file simultaneously.

ashwin-ant

[Repo] modelcontextprotocol/registry: A community driven registry service for Model Context Protocol (MCP) servers.

The modelcontextprotocol/registry is a Go-based community-driven registry service for Model Context Protocol (MCP) servers, enabling discovery and management of MCP server implementations. With 6,518 stars, it serves as a central hub for the MCP ecosystem, allowing developers to register, discover, and integrate MCP servers.

Berkeley LLM Function-Calling Leaderboard and State-of-Art OpenFunctions-v2

Berkeley's LLM Function-Calling Leaderboard evaluates large language models on their ability to accurately call functions and APIs. OpenFunctions-v2 represents the state-of-the-art model for function-calling tasks, demonstrating improved performance in understanding and executing function invocations.

shishirpatil

AWS will offer HF’s products to its customers and run its next LLM tool

AWS has partnered with Hugging Face to offer its AI products and services to AWS customers, including hosting Hugging Face's next-generation LLM tool. This collaboration represents AWS's strategic move to compete in the rapidly growing generative AI market alongside other major cloud providers.

mmq

Don't trust AI agents

This article discusses security concerns and the importance of not blindly trusting AI agents. It appears to be from Nanoclaw's blog and likely covers their security model and best practices for safely deploying AI agents.

gronky_

Signal leaders warn agentic AI is an insecure, unreliable surveillance risk

Signal's leadership warns that agentic AI systems pose significant security, reliability, and privacy risks, particularly regarding surveillance capabilities. The article highlights concerns about autonomous AI agents operating without sufficient safeguards and oversight mechanisms.

speckx

Improving GPT 4 Function Calling Accuracy

This article discusses strategies for improving the accuracy of GPT-4's function calling capabilities, which is essential for building reliable AI agents that can execute specific tasks and integrate with external tools. The content covers best practices and techniques to enhance how GPT-4 interprets and calls functions correctly.

hudixt

[Release] openclaw/openclaw v2026.2.24-beta.1: openclaw 2026.2.24-beta.1

OpenClaw v2026.2.24-beta.1 release introduces enhanced auto-reply/abort shortcuts with multilingual support, Android native onboarding with new tab-based navigation, and provider-agnostic Talk configuration. The update includes critical security and routing fixes for cross-channel session isolation, heartbeat delivery improvements, and extensive channel-specific reliability enhancements across Discord, WhatsApp, Matrix, and Telegram.

steipete

Claude 4

The provided content only contains a title 'Claude 4' and a URL reference to Anthropic's news page, with no actual article content or technical information to summarize.

meetpateltech

[Release] anthropics/claude-code v2.1.63: v2.1.63

Claude Code v2.1.63 release includes new slash commands (/simplify, /batch), HTTP hooks support, and numerous bug fixes and memory leak resolutions. Key improvements address MCP server handling, git worktree configuration sharing, message ordering in REPL bridge, and long-running session stability.

ashwin-ant

[Release] crewaiInc/crewAI 1.10.1a1: 1.10.1a1

crewAI v1.10.1a1 introduces asynchronous invocation support in step callback methods and implements lazy loading for heavy dependencies in the Memory module. The release includes documentation updates, refactoring improvements, and bug fixes for release notes.

greysonlalonde

[Release] anthropics/claude-code v2.1.68: v2.1.68

Claude Code v2.1.68 updates Opus 4.6 to default to medium effort for Max and Team subscribers, reintroduces the 'ultrathink' keyword for enabling high effort, and removes Opus 4 and 4.1 from the first-party API with automatic migration to Opus 4.6.

ashwin-ant

[Release] anthropics/claude-code v2.1.66: v2.1.66

Claude Code v2.1.66 release from Anthropics. This is a version update for the claude-code tool, likely containing bug fixes, improvements, or new features compared to previous versions.

ashwin-ant

Show HN: MCP server for searching and downloading documents from Anna's Archive

An open-source MCP server that integrates Anna's Archive with Claude Desktop, enabling users to search for and download books directly through the Claude interface. The project includes automated builds via GitHub Actions and comprehensive documentation for installation and configuration.

iosifache

[Release] google/adk-python v1.22.1: v1.22.1

Google ADK Python v1.22.1 release includes bug fixes for the `adk migrate session` CLI command that was previously removed, and fixes escaping of database reserved keywords to prevent SQL errors.

wuliang229

[Release] microsoft/autogen python-v0.7.5: python-v0.7.5

Microsoft AutoGen releases python-v0.7.5 with 18 bug fixes and improvements including support for linear memory in RedisMemory, thinking mode support for Anthropic client, fixes for streaming response handling, GraphFlow cycle detection, and security enhancements with DockerCommandLineCodeExecutor. The release addresses issues with Bedrock streaming, JSON schema conversion, token counting, and MCP session management.

ekzhu

[Release] anthropics/claude-code v2.1.59: v2.1.59

Claude Code v2.1.59 introduces auto-memory context saving with /memory management, a new /copy command for interactive code block selection, and improved bash command prefix suggestions for compound commands. The release also includes memory optimization for multi-agent sessions, fixes for MCP OAuth token refresh race conditions, and better error messaging for deleted working directories.

ashwin-ant

[Release] microsoft/autogen python-v0.6.1: python-v0.6.1

Microsoft AutoGen released python-v0.6.1 with bug fixes for GraphFlow cycle checking and graph validation logic. The release also adds function calls and results tracking in ToolCallSummaryMessage.

ekzhu

LLMCompiler: An LLM Compiler for Parallel Function Calling

LLMCompiler is a compiler framework that enables large language models to execute multiple function calls in parallel, improving efficiency and reducing latency in LLM-based applications. It optimizes function calling by analyzing dependencies and scheduling independent operations concurrently.

ngaut

Show HN: Deepmark AI- LLM assessment tool for task-specific metrics on your data

Deepmark AI is an LLM assessment tool that evaluates language models using task-specific metrics on custom datasets. It enables developers to benchmark and compare LLM performance across different use cases and data domains.

Vasco2006