how to run long-lived claude code agents from your phone

说明与用途

从手机运行长期Claude Code代理需要保持主机唤醒、使用远程控制，并将状态存储在git支持的长期记忆中，以防止上下文退化。

问题

Running Claude Code agents for extended periods while steering them from a phone requires keeping the host awake, maintaining healthy sessions, and preventing context degradation.

结果

Agents ran continuously for weeks, controlled from a phone over satellite, with context maintained via aigent-os and host kept awake by OS commands.

工作流

• 1. Ensure a Pro, Max, Team, or Enterprise plan and log in via claude.ai.
• 2. Run "claude remote-control" in the agent’s project directory or use "/remote-control" in an existing session.
• 3. Connect from the Claude app (scan QR or select session).
• 4. Keep the host computer awake using OS-specific commands (e.g., macOS: caffeinate -dimsu; Linux: systemd-inhibit --what=idle:sleep --why="agents running" sleep infinity; Windows: powercfg /change standby-timeout-ac 0 /change hibernate-timeout-ac 0).
• 5. Run sessions inside tmux, plug in power, disable auto-restarts, and test remote control before leaving.
• 6. Use aigent-os to periodically save a context capsule, clear the session with "/clear", and resume from the capsule plus git-backed long-term memory to avoid context window exhaustion.

证据

i ran my newsletter business from my phone for 14+ days in september while elk hunting. phone, Starlink Mini, no laptop. the agents ran on my computer at home and i steered them from the Claude app.

来源

the loop is simple: 1) the agent saves a context capsule; 2) it runs /clear; 3) it resumes from the capsule plus its long-term memory, which is git-backed.

来源

for multiple agents, i run one session per agent, each in its own directory, so each shows up as its own session in the app.

来源