Commandline - 2

2026 (11)

June (7)

commandline: Microsoft Scout: From personal project to enterprise-ready personal agent

Microsoft Command Line Blog

2026-06-02

Microsoft Scout is described as an always-on personal AI agent that moved from a personal project into an enterprise-ready, secure product. The article is a discussion with Omar Shahine and Jakob Werner about the design and build process.

commandline: The agent optimization loop and how we built it in Foundry

Microsoft Command Line Blog

2026-06-02

The article describes Microsoft’s approach to agent optimization in Foundry, reframing the problem to identify the factors that matter most and implementing a new optimization loop. The provided excerpt is high level and does not include technical details.

commandline: Grounding at scale: Engineering the retrieval system for the agentic web

Microsoft Command Line Blog

2026-06-02

Microsoft introduced Web IQ, a new grounding system for the agentic web. The note is brief and does not include implementation details or feature specifics.

commandline: Turn specs into evals for any agent with ASSERT

Microsoft Command Line Blog

2026-06-02

ASSERT is an open-source framework that turns natural-language behavior requirements into executable evaluations for AI models and agents. It is designed for evaluation and regression testing from written specs.

commandline: Introducing Agent Control Specification: Portable runtime governance for AI Agents

Microsoft Command Line Blog

2026-06-02

ACS is presented as an open, vendor-neutral standard for applying runtime governance across the AI agent lifecycle. It is designed to work independently of the agent framework, runtime, or policy engine.

commandline: Composing a new platform for agent-first devices

Microsoft Command Line Blog

2026-06-02

The article states that new interaction technology will enable new categories of computers. No additional technical details are provided in the excerpt.

commandline: Five ways we’re confusing AI capability and AI reality—and how to bridge the gap

Microsoft Command Line Blog

2026-06-02

Microsoft partnered with Lectures on Tap for Build on Tap, where CTO Kevin Scott discussed a perceived gap between AI capability and how AI is actually being used. The piece frames this as a disconnect between industry expectations and operational reality.

May (4)

commandline: Disposable agents, durable memory: The architecture behind Squad

Microsoft Command Line Blog

2026-05-27

The article describes Squad, a file-backed agent-team architecture built around human oversight. It emphasizes inspectable memory, explicit orchestration, and moving governance rules from prompts into code.

commandline: Inside the Microsoft Agent Framework: How we designed a layered SDK

Microsoft Command Line Blog

2026-05-27

Microsoft describes the Agent Framework as a layered SDK that supports different agent architectures. It highlights agent loops, workflows, and harnesses as the main building blocks for both simple assistants and more complex multi-agent systems.

commandline: Introducing Microsoft’s EngThrive framework: Understanding developer productivity in the agentic AI era

Microsoft Command Line Blog

2026-05-27

Microsoft introduced EngThrive, a framework for measuring developer productivity in the agentic AI era. It evaluates engineering work across Speed, Ease, Quality, and Thriving.

commandline: How Excel got agentic: An interview with Microsoft Director of Science Mukul Singh

Microsoft Command Line Blog

2026-05-27

Microsoft says it used two years of research to add agentic capabilities to Excel, based on an interview with Director of Science Mukul Singh. The piece positions this work as part of bringing research into product, with implications beyond Excel.