Agent Ops: Editor merges, Astra, routing, orchestration, audits

Visual Studio Code 1.136 introduces agent merges for pull requests. VS Code previews Agent Merge so agents can address reviews, fix failing checks, and finish pull requests, with multi-root agent chats and privacy controls. Treat editors as a control plane—this changes review ownership, CI handoffs, and how you gate agentic pushes (Principles 03,06,15).

GPT-6 Astra on OpenRouter. GPT-6 Astra surfaces as OpenAI’s most capable agentic model for long-horizon, browser-and-computer-enabled engineering and research workflows. Having Astra reachable through OpenRouter shifts deployment and model-selection trade-offs—plan for higher capability endpoints when you design long-running agents (Principles 06,09).

Portal by Spotify cut my Claude Code token usage by 90%. Portal’s AiKA Modes route boilerplate and heavy I/O to cheaper models, reducing token use dramatically. Model routing is an architectural primitive for outcome engineers—use it to balance cost, latency, and capability across agent pipelines (Principles 06,07).

OpenClaw Power, MacBook Simplicity: Five Days With Grok Bot. Grok Bot makes multi-agent setups frictionless, turning bots into English-programmable building blocks instead of brittle, user-owned platforms. If you’re building agent ecosystems, orchestration and agent ergonomics become first-class engineering problems to solve (Principles 01,03,09).

OpenAI says it is building a framework for reporting misalignment incidents during training, evaluation, and deployment. OpenAI is building an incident-reporting framework to surface, track, and audit misalignment and failure modes across the model lifecycle. Outcome engineering needs these pipelines—plan for incident telemetry, reporting workflows, and auditable trails as part of any production agent deployment (Principles 10,16).