Passing Once Isn't Reliable — This Week's Agent Engineering Puts the Harness Before the Model

This digest covers AI agent developments from 2026-08-18 to 2026-08-25: orchestration patterns, tool/function calling, memory, planning loops, multi-agent coordination, and agent evaluation.

AgentWeave: Routing Before Reasoning for Efficient Function Calling in Tool-Rich Language Models — cuts tool exposure 70%, latency 51%.

One Success Isn't Reliability: Thinkingbox, a Sandbox and Benchmark for Agents in Stateful Business Workflows — pass@1 success hides a pass@20 collapse.

Spine-Branch Coordination for Multi-agent Computer Use — solves VM state merging across parallel agents.

Introducing LangSmith Tuned Evaluators, starting with Perceived Error — grades 100% of production traffic, cheaply.