Anthropic's New Rules of Context Engineering for Claude 5 Generation Models
Anthropic's official guide to context engineering for Claude 5 models — covering memory, long-horizon tasks, and system prompt simplification.
How-Tos
525 piecesChannelGuard: Safe Models Don't Compose Into Safe Multi-Agent Systems
New arXiv research shows safe LLMs can compose into unsafe MAS — ChannelGuard offers training-free inter-agent channel auditing with IB gates.
Claude Managed Agents Platform Update: Effort Levels, Seed Sessions, and Expanded Webhooks
Claude Managed Agents gains effort-level control, seed sessions with up to 50 initial events, and expanded lifecycle webhooks for production agentic workflows.
Claude Voice Mode Now Runs on Opus, Sonnet, and Haiku with Gmail, Slack, and Notion Integrations
Claude voice mode now runs on Opus, Sonnet, and Haiku with Gmail, Slack, and Notion integrations — a major step toward voice-driven agentic workflows.
Claude Code v2.1.219: Nested Subagents Up to Depth 3, Opus 5 Default, and Dynamic Workflow Sizing
Claude Code v2.1.219 enables 3-tier nested subagent hierarchies and ships Opus 5 as default — a foundational shift for autonomous pipelines.
DeepSeek V4 Reaches General Availability — MIT License, 1.6T MoE, SWE-bench 80.6%, Ultra-Cheap Agentic Pricing
DeepSeek V4 hits full GA with MIT license, 80.6% SWE-bench, and pricing as low as $0.14/M tokens — a major option for budget-conscious agentic builders.
Kimi K3 Has a ~51% Hallucination Rate Moonshot Didn't Publish — What Agentic Teams Need to Know Before July 27
Independent testing reveals Kimi K3's hallucination rate rose from 39% to ~51% before its July 27 open-weights drop — critical data for agentic production use.
How to Run Visa's VVAH Agentic Security Scanner on Your Codebase
Visa open-sources VVAH under Apache 2.0 — an 11-stage Claude-powered agentic SAST pipeline that goes from AI-discovered exploit to production-ready fix.
How to Give Your OpenClaw Agents Real Work Memory with Screenpipe
Set up Screenpipe to give your OpenClaw agents searchable memory of your real screen and audio activity via its MCP-compatible API on port 3030.
Echo: Open-Weight Ensemble Model Router Matches Fable Quality at 1/3 the Cost
Echo from Tracer routes across GLM-5.2, Kimi K2.7 and others dynamically, matching frontier quality at 1/3 the inference cost via one OpenAI-compatible endpoint.