The board · maintained by hand

Harnesses

What runs the loop. The programs that hold a model, its tools and its permissions together for long enough to finish a job.

Claude Code Anthropic Terminal, desktop, web and IDE. The reference implementation of a coding agent with real permissions. coding

Claude Agent SDK Anthropic Claude Code's loop, tools and context management as a library. Python and TypeScript only — other languages drive the CLI as a subprocess. library

Managed Agents Anthropic Hosted REST API for long-running agents. Anthropic runs the agent and the sandbox, so you don't operate session infrastructure. hosted

Codex CLI OpenAI Sandboxed by default and open source. The usual pick when those two are hard requirements. coding

Gemini CLI Google Apache-2.0, strong on whole-repository context. The free tier ended 18 Jun 2026, so it's now a budgeted decision. coding

Editor agents Cursor, Zed, Copilot, others Agent loops living inside the editor. Tighter feedback, narrower blast radius. in-editor

Hand-maintained, not a feed. Checked 12 Aug 2026 against Anthropic's model docs, OpenAI's model list and Google's Gemini docs. Vendor facts come from those pages; the read on each harness is judgement, not vendor copy.