Tool Use

OpenAI Agents Python v0.13.4 is a patch release that fixes AnyLLM response replay input validation by sanitizing inputs before validation. The release includes documentation updates for extension feature mapping and routine release automation tasks.

Claude Code v2.1.89 is a major release introducing deferred permission decisions for headless sessions, flicker-free rendering, improved MCP connection handling, and numerous bug fixes across tool execution, file handling, and UI rendering. Key improvements include named subagents, enhanced auto-mode permission workflows, and fixes for critical issues like memory leaks, LSP crashes, and prompt cache misses in long sessions.

crewAI v1.13.0a3 introduces token usage tracking in LLM events and tool metadata publishing to AMP. The release includes bug fixes for GPT-5.x model compatibility, comprehensive documentation updates on SSO configuration and agent capabilities, and significant refactoring converting Flow and LLM classes to Pydantic BaseModel with improved type annotations.

OpenClaw v2026.3.31 is a major release introducing significant breaking changes to node execution, plugin SDK deprecations, and stricter security defaults for dangerous code scanning. Key improvements include a unified background task control plane, new channel integrations (QQ Bot), enhanced Matrix/MCP support, and security hardening across gateway authentication and node command execution. The release emphasizes safer defaults while expanding platform capabilities for agents, channels, and integrations.

OpenClaw v2026.3.31-beta.1 introduces significant breaking changes to node execution, plugin SDK deprecation, and stricter security defaults for plugin/skill installation. Key improvements include unified background task management with flow control, expanded channel support (QQ Bot), enhanced MCP capabilities with remote HTTP/SSE servers, and security hardening across gateway authentication, node commands, and event handling. Multiple channel-specific enhancements (Matrix, LINE, Teams, WhatsApp, Slack) and provider integrations (OpenAI Responses, Pi Codex, TTS diagnostics) round out the release.

LangChain v1.2.14 release includes security patches, performance improvements, and bug fixes. Key updates address CVE-2026-4539 via Pygments dependency upgrade, reduce initialization speed by 15%, fix recursion limit handling in create_agent, and improve token counter recognition for ChatAnthropicVertex. The release also updates dependencies, adds async implementations, and enhances Azure AI Foundry compatibility.

This cookbook teaches context engineering strategies for long-running AI agents, focusing on three key techniques: compaction (summarizing context), tool-result clearing (removing re-fetchable tool outputs), and memory (persistent external storage). The guide addresses context rot—the degradation of model performance as context windows grow—and provides practical implementations using Claude's API. Through a research agent example, it demonstrates how to combine these strategies to manage token growth, maintain conversation continuity, and persist knowledge across sessions.

This guide teaches how to build knowledge graphs from unstructured text using Claude's structured outputs for entity extraction, relation mining, and entity resolution. Rather than training traditional NER and relation classifiers, Claude handles each stage via prompts, enabling multi-hop graph reasoning without a database. The approach uses Haiku for high-volume extraction and Sonnet for entity resolution, with techniques transferable to production databases like Neo4j or PostgreSQL.

OpenAI Agents Python v0.13.3 is a maintenance release addressing two critical bugs: fixing stale hydrated input IDs in server conversation tracking and enabling raw image_url content parts in chat completions. The release also includes documentation improvements for non-OpenAI providers, updated translations, and code quality enhancements with Ruff linting.

Claude Code v2.1.88 is a maintenance release focused on stability, performance, and bug fixes across rendering, tool execution, and session management. Key improvements include flicker-free rendering, prompt cache fixes for long sessions, structured output schema reliability, and numerous UI/UX enhancements. The release addresses critical issues like memory leaks, file handling on Windows, voice mode permissions, and terminal compatibility across platforms.

langchain-openrouter version 0.2.1 was released with a critical fix for passing attribution headers through httpx default_headers instead of the previous method. The release includes updated model profile data and a dependency bump for the requests library from 2.32.5 to 2.33.0. This patch addresses header handling in the OpenRouter integration for LangChain.

OpenClaw v2026.3.28 is a major release featuring breaking changes to Qwen OAuth migration and config validation, alongside significant enhancements including xAI Responses API integration, MiniMax image generation, async tool approval hooks, ACP channel bindings, and unified file upload actions across platforms. The release includes numerous bug fixes for Telegram splitting, WhatsApp echo loops, Mistral API compatibility, and security improvements for credential auditing.

OpenClaw v2026.3.28-beta.1 introduces significant breaking changes including removal of deprecated Qwen OAuth integration and stricter config validation, while adding major features like async tool approval hooks, MiniMax image generation, xAI search integration, and unified file upload actions across platforms. The release includes numerous fixes for Telegram message splitting, WhatsApp echo loops, image tool fallbacks, and security improvements for credential auditing.

Claude Code v2.1.86 is a maintenance release focused on stability, performance, and bug fixes. Key improvements include session tracking via headers, VCS exclusion enhancements, fixes for file operations outside project root, memory/performance optimizations, and corrections to UI interactions and plugin handling. The release also improves token efficiency through prompt caching and reduced overhead in file mentions.

LangChain Core version 1.2.23 was released with minimal changes from 1.2.22. The release includes a revert of a trace invocation params metadata fix and a dependency bump for the requests library from 2.32.5 to 2.33.0. This appears to be a patch release addressing a regression in the previous version.

Claude Code v2.1.85 introduces MCP server flexibility with environment variables for multi-server support, conditional hook execution to reduce overhead, and improved OAuth compliance with RFC 9728. The release includes significant bug fixes for context management, plugin handling, terminal compatibility, and streaming performance, plus enhancements to file autocomplete and scroll performance on large transcripts.

OpenAI Agents Python v0.13.2 is a maintenance release focusing on bug fixes and security improvements. Key updates include fixing docstring parsing with griffelib, preventing private tool metadata leakage in persisted sessions, ensuring LiteLLM reasoning_effort portability across providers, and pinning LiteLLM to version 1.82.6 to mitigate supply chain risks. The release also adds external web access capability to WebSearchTool and updates default reasoning effort for newer models.

langchain-exa version 1.1.0 introduces a significant change to the default search type from neural to auto, along with dependency updates and maintenance improvements. The release includes bumps to core dependencies like langchain-core (minimum 1.2.21), requests, orjson, and langsmith, ensuring compatibility and security. Various CI/CD improvements and code quality enhancements were made, including typing targets and relative import restrictions across packages.

CrewAI 1.12.0 introduces significant enhancements including Qdrant Edge storage backend, native OpenAI-compatible provider support (OpenRouter, DeepSeek, Ollama, vLLM, Cerebras, Dashscope), agent skills implementation, and automatic root_scope for hierarchical memory isolation. The release adds Arabic language support, a new docs-check CLI command, logout functionality, and resolves multiple bugs across type checking, HITL flow system, and memory management. Documentation improvements include a new CONTRIBUTING.md guide and instructions for using CrewAI without LiteLLM.

Claude Code v2.1.84 introduces PowerShell tool support for Windows, enhanced environment variable controls for model configuration, new hooks for task and worktree creation, and numerous bug fixes and performance improvements. Key additions include idle timeout configuration, channel plugin allowlists, deep link improvements, and better MCP server handling. The release also addresses critical issues with voice input, workflow subagents, and terminal rendering while improving startup performance and prompt cache efficiency.