Skills & Tools

Google ADK Python v1.29.0 introduces significant enhancements to authentication, security, and tooling capabilities. Key additions include auth scheme/credential support for MCP toolsets, configurable resource limits and shell metacharacter blocking in BashTool, new EnvironmentToolset for file I/O and command execution, and improved BigQuery logging with configurable view prefixes. The release also adds support for model endpoints in Agent Registry, regional endpoints for SecretManager, and enhanced session management with custom session IDs and shallow-copy options.

This tutorial demonstrates building a webhook-triggered SRE incident response agent using Claude Managed Agents that automatically investigates production alerts, consults runbooks, proposes infrastructure fixes via pull requests, and gates merging behind human approval. The agent combines built-in sandbox tools (bash, read, edit) with custom tools for PR management and human-in-the-loop approval, providing complete audit trails in the Anthropic Console. The example uses mocked PagerDuty, GitHub, and Datadog integrations to focus on agent patterns, with guidance for swapping in real services.

This tutorial introduces the Claude Managed Agents API by walking through a practical debugging workflow where an agent iteratively fixes failing tests in a Python package. It covers the three core resources (Agent, Environment, Session), file mounting, and the streaming event loop pattern. The example demonstrates how agents autonomously discover the iterate-observe-fix loop by running tests, reading failures, editing code, and repeating until all assertions pass.

This cookbook demonstrates building a Slack data analyst bot using Claude Managed Agents and Bolt for Python. Users mention the bot with a CSV file to receive narrative analysis reports in-thread, with support for multi-turn follow-ups within the same session. The implementation handles file uploads, streams agent progress updates, and manages session persistence across Slack threads.

Claude Code v2.1.97 introduces a focus view toggle for better prompt visibility, enhanced status line customization with git worktree support, improved permissions handling with stricter Bash tool checks, and numerous bug fixes across MCP connections, resume functionality, NO_FLICKER mode rendering, and authentication. The release also improves auto-approval logic for safe commands, image compression consistency, and CJK text input handling.

crewAI 1.14.1rc1 introduces async checkpoint TUI browser and streaming output enhancements, along with critical security updates and architectural improvements. Key changes include CVE-2026-1839 resolution via transformers bump, dynamic tool field exclusion, and refactored provider architecture using BaseModel discriminators. The release focuses on async/streaming capabilities, security hardening, and code quality improvements across the framework.

crewAI 1.14.1 introduces async checkpoint TUI browser functionality and improves streaming output handling with async context managers. The release includes critical security updates (CVE-2026-1839 fix via transformers bump), bug fixes for checkpoint registration and tool name sanitization, and architectural refactoring to use dynamic field exclusion and shared singleton patterns. Key improvements enhance developer experience through better CLI tooling and provider type discrimination.

The Claude Agent SDK provides built-in session management functions to build conversation history sidebars without writing custom parsers. This cookbook demonstrates how to list, read, rename, tag, and fork sessions stored as JSONL transcripts on disk. By leveraging SDK primitives like list_sessions(), get_session_messages(), and rename_session(), developers can implement session browsers for desktop apps, IDE extensions, and chatbots without managing file I/O directly.

This cookbook demonstrates building a Claude-powered threat intelligence enrichment agent that autonomously investigates Indicators of Compromise (IOCs) by querying multiple threat intel sources, correlating findings, mapping to MITRE ATT&CK, and generating structured reports for SIEM/SOAR integration. The agent uses Claude's tool-use capabilities to decide which intelligence sources to query, chain tool calls based on discoveries, and convert free-text analysis into analyst-ready JSON reports. The architecture uses simulated threat intel backends that can be swapped with real APIs (VirusTotal, AbuseIPDB, Shodan, etc.) without changing orchestration logic.

OpenClaw v2026.4.5 is a major release introducing breaking config changes with migration support, new media generation tools (video, music, image), expanded provider integrations (Qwen, Fireworks AI, StepFun, Amazon Bedrock Mantle), multilingual control UI support, enhanced approval workflows for iOS and Matrix, experimental memory dreaming features with weighted recall, and improved prompt caching stability across multiple dimensions.

Claude Code v2.1.91 introduces MCP tool result persistence overrides supporting up to 500K characters, adds shell execution disabling options, enables multi-line prompts in deep links, and allows plugins to ship executables. The release includes critical fixes for transcript chain breaks, terminal key handling, plan mode persistence in remote sessions, and JSON schema validation, plus performance improvements and enhanced API guidance.

CrewAI v1.13.0 introduces significant architectural improvements including RuntimeState RootModel for unified serialization, enhanced event telemetry for skills and memory, and A2UI extension support. The release includes critical bug fixes for GPT-5.x model compatibility, RBAC permission alignment, and Windows lancedb compatibility. Performance optimizations reduce framework overhead through lazy event bus implementation and conditional tracing.

OpenClaw v2026.4.1 is a maintenance and feature release introducing chat-native task boards, SearXNG web search integration, Amazon Bedrock Guardrails support, and enhanced channel features (Feishu comments, WhatsApp reactions, Telegram improvements). The release includes significant fixes for task management, execution approvals, session handling, and plugin runtime dependencies, with improved error handling and configuration flexibility across gateways, channels, and agent systems.

OpenClaw v2026.4.1-beta.1 introduces significant enhancements across chat, web search, cloud integrations, and channel management. Key additions include a native `/tasks` background task board, SearXNG web search provider, Amazon Bedrock Guardrails support, and improved Feishu document collaboration. The release also addresses critical fixes for task management, execution approvals, session routing, and channel-specific issues to improve stability and user experience.

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.

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.