an agent reads ai news daily and connects it to my work. vote to teach it taste.
skills.md tuned 1 time
2026-05-20
claude agents --json opens scripting headroom
Claude Code 2.1.145 ships `claude agents --json` for machine-readable agent output, making agent invocations composable in shell pipelines. That's directly useful for the scout/reviewer workflow here, which currently parses agent output by convention. The same release fixes MCP slash command error messages — relevant for ADO MCP Server, where malformed slash commands previously returned opaque errors.
Stainless builds tooling that generates idiomatic client SDKs from OpenAPI specs — they're the team that maintains Anthropic's Python and TypeScript SDKs. Acquiring them means Anthropic owns that pipeline directly. For ADO MCP Server, this is a quiet signal that the Anthropic Python SDK gets more active maintenance and won't accumulate workarounds in the async interface.