Leading the Claude and Codex desktop apps phase by phase, instead of the CLI

说明与用途

OpenDaemon 让 AI 编码代理在 Claude、Codex 和 Antigravity 桌面应用中按阶段执行，提供可视化检查和人工批准，提升用户信任并减少误报完成。

问题

Automating coding agents typically relies on CLI scripts, which hides agent activity from users and reduces trust.

结果

By moving agent control into desktop apps, users gain visibility of each phase, reduce false positives of completion, and maintain control over approvals.

工作流

• Plan first: describe the app and generate phases and checklist; no code written before plan.
• Brief each phase: OpenDaemon writes a full brief and starts a fresh conversation in the coding app.
• Check, don’t trust: agent reads report and screenshots; phases that fail to open or run are retried.
• Playtest last: final phase runs built code, walks checklist, saves screenshots; unverified items are listed.
• You stay in charge: approve each phase or the entire plan with full trust.

证据

When people automate coding agents, they usually script the CLI. I built OpenDaemon to lead the agent inside the desktop apps instead

来源

Plan first. You describe the app once. A planner writes phases plus a checklist: what you asked for, plus what anyone would expect from that kind of thing. No code is written before the plan exists.

来源

Check, don’t trust. It reads the report and screenshots. A phase that doesn’t open or run goes back to be fixed, never marked done.

来源

"Done" is the hard problem. Agents report success eagerly. Requiring something you can open at the end of every phase removed most false "done"s.

来源