Harness overview

Harnesses are the reason ctrlyoke can keep one workflow model while talking to very different AI assistants.

See the generated harness compatibility matrix for capability fidelity, evidence, version scope, caveats, and remediation guidance.

At a glance

Harness	Runs where	Best for	Notes
copilot-cli	CLI	Default path for GitHub users	Good general-purpose default
claude-cli	CLI	Long-form reasoning and coding	Strong session continuity
codex-cli	CLI	OpenAI Codex workflows	Sandbox/approval modes, not tool patterns
opencode-cli	CLI	Provider-flexible OpenCode setups	Useful when OpenCode is already your standard
qwen-cli	CLI	Qwen Code users	No image attachments
antigravity-cli	CLI	Google Antigravity workflows	Uses agy; coarse permission modes only

Choosing a starting harness

• Start with copilot-cli if you want the path ctrlyoke is built around by default.
• Use claude-cli if your team already works in Claude Code.
• Reach for the other CLI harnesses when they already match your preferred assistant or hosting model.

All harnesses use the same workflow schema and share the core runtime: interactive terminal mode, headless execution, per-harness additionalArgs, CLI session IDs, MCP server attachment, and portable agents and skills. What differs is enforcement of tool permission lists, which hook events fire, per-server MCP tool filtering, and image attachments — check the compatibility matrix before relying on one of those.

See the per-harness pages for setup notes and tradeoffs.