Cloudflare Clef and Strands Decider: Decision Models for AI Agents Explained
Cloudflare Clef (27B, 9B Flash) and Amazon Strands Decider 2B are small Apache 2.0 decision models. How they work, where they fit in agent loops, and cost.
OpenAI Agents API Public Beta: Managed Sandboxes vs Self-Hosted Agent Runtimes
OpenAI Agents API entered public beta with sandboxes near $0.03 per 20 minutes. Managed agent runtimes vs self-hosted sandboxes on cost and control.
Meta Muse and Muse Code: Personal Agent Architecture and Pricing Explained
Meta Muse is a 24/7 personal agent and Muse Code a coding tier. What is confirmed on tiers and quota, plus engineering trade-offs of always-on agents.
AI Agent Audit Logging and Identity: Lessons From the OpenAI Agent Access Incident
OpenAI agents reportedly accessed 55 sites with obscured audit logs. How to design agent identity, scoped credentials, and tamper-evident audit trails.
AI Agent Evaluation Harness: Trajectory Evals in Production
How to evaluate tool-using AI agents: trajectory vs outcome evals, simulated environments, LLM judges, regression gates in CI, with a reference architecture.
MCP Goes Stateless: Migrating to the 2026-07-28 Spec
MCP's 2026-07-28 spec removes protocol sessions and the handshake, going fully stateless - a migration guide to what breaks and how to fix it.
langchain-mcp-adapters vs Native langchain.mcp (2026)
LangChain 1.4 folds MCP support into langchain.mcp, replacing langchain-mcp-adapters. Migration guide, the MCPAdapter class, elicitation, and OAuth 2.1 setup.
MCP Tasks vs Streaming vs Webhooks: Handling Long-Running Agent Tool Calls
The MCP Tasks extension turns long tool calls into a polling state machine. How it compares to SSE streaming and webhooks, and when each one breaks.
OPC UA Companion Specifications as MCP Tools: Wiring Industrial Semantics into AI Agents (2026)
OPC UA has 400+ companion specifications of industrial semantics. How to expose them to AI agents as MCP tools and RAG corpora - architecture, pitfalls and safety.
AI Agent Sandboxes Compared: Firecracker vs gVisor vs Kata for Untrusted Code in 2026
Running LLM-generated code safely in 2026: Firecracker, gVisor and Kata compared on isolation boundary, cold start, GPU passthrough and operational cost.