Field guide to the AI agents colonizing your terminal

What it does This is a curated “awesome list” that catalogs more than 110 CLI coding agents—AI tools that live in your terminal, read repos, edit files, and shell out commands—alongside the harnesses used to orchestrate, sandbox, or run them in parallel. Think of it as a phone book for the current swarm of terminal-native pair programmers.

The interesting bit Rather than dumping every AI tool into one bucket, the list splits agents from their infrastructure. It explicitly calls out session managers, parallel runners, and autonomous loops as first-class categories, which is where the real complexity lives once you stop demoing and start shipping.

Key highlights

• Covers open-source agents (Aider, Goose, OpenCode), platform-backed tools (Claude Code, Codex CLI, Gemini CLI), and the emerging “OpenClaw” ecosystem.
• Organizes entries by license and backing, with separate sections for open-source and closed-source terminal agents.
• Includes dedicated harness and orchestration categories: session managers, parallel runners, autonomous loops, and agent infrastructure.
• Offers a clear working definition of what qualifies as a CLI coding agent—terminal-native, with direct filesystem and shell access—as distinct from chat-based assistants.

Caveats

• The README is strictly a directory: expect curation and categorization, not hands-on benchmarks or comparative reviews.
• The document carries a “last updated” date in 2026 and references future events like a “March 2026 source leak,” so some entries appear speculative or forward-dated.

Verdict Worth a bookmark if you are trying to pick a terminal agent or orchestration harness and want a lay of the land. Skip it if you are looking for hands-on tutorials or performance shootouts; this is a map, not a manual.