Agent skill

restart-claude

Use when the user asks to restart Claude Code, or after changing settings/hooks/keybindings.

Stars 1,038
Forks 113

Install this agent skill to your Project

npx add-skill https://github.com/zylos-ai/zylos-core/tree/main/skills/restart-claude

SKILL.md

Restart Claude Code Skill

Restart Claude Code session - sends /exit and lets activity-monitor daemon handle the restart.

When to Use

  • After changing Claude Code settings, hooks, or keybindings
  • When Claude needs to reload configuration
  • To clear temporary state without upgrading
  • User explicitly asks to restart

Pre-Restart Checklist

Before sending /exit, complete these steps in order:

1. Stop background tasks

Check for running background agents (Task tool). If any are active, stop them to avoid orphaned work.

2. Sync memory

Update memory files (state.md, sessions/current.md, etc.) to preserve important context that would otherwise be lost on restart.

3. Write a session handoff summary

Write a brief message covering:

  • What was being worked on (active tasks, user requests in progress)
  • Current state (what's done, what's pending, any blockers)
  • What the next session should pick up (if anything)

4. Send the handoff summary

Determine who to notify:

  • If actively collaborating with a user: Send the summary to that user's channel via C4 (their reply via path). This keeps the user informed AND records the context into C4 conversation history.
  • If no active user conversation: Send the summary to the web console channel. This still records it into C4 so the new session's startup hook (c4-session-init) will include it in the conversation context.

The goal is twofold: (a) the user knows what's happening, and (b) the handoff summary appears in C4 conversation history, so the new session can seamlessly continue the work.

5. Enqueue /exit

bash
node ~/zylos/.claude/skills/comm-bridge/scripts/c4-control.js enqueue --content "/exit" --priority 1 --block-queue-until-idle

How It Works

  1. Enqueue /exit: Puts /exit into the control queue (priority=1, block_queue_until_idle)
  2. Block subsequent messages: block_queue_until_idle prevents other messages from being dispatched
  3. Deliver when idle: Dispatcher delivers /exit to tmux when Claude is idle
  4. Daemon restart: activity-monitor detects exit and restarts Claude

Expand your agent's capabilities with these related and highly-rated skills.

zylos-ai/zylos-core

web-console

Built-in web interface for communicating with Claude without external services. Use when setting up or configuring the web console channel, or troubleshooting browser-based access.

1,038 113
Explore
zylos-ai/zylos-core

component-management

Guidelines for managing zylos components via CLI and C4 channels. Use when installing, upgrading, or uninstalling components, or when user asks about available components.

1,038 113
Explore
zylos-ai/zylos-core

comm-bridge

C4 communication bridge — central gateway for ALL external communication (Telegram, Lark, etc.). Use when replying to users via the "reply via" path, sending proactive messages to external channels, querying recent conversations or checkpoint status (prefer c4-db.js CLI; sqlite3 OK for unsupported queries), fetching conversation history for Memory Sync, or creating checkpoints after sync. Incoming messages are queued by channel bots and delivered to Claude via a PM2 dispatcher daemon. Session-start hooks automatically provide conversation context and can trigger Memory Sync when unsummarized conversations exceed the configured threshold.

1,038 113
Explore
zylos-ai/zylos-core

new-session

Start a new session when context is high. Claude uses /clear, Codex uses /exit. Use when context is high or when a fresh session is needed.

1,038 113
Explore
zylos-ai/zylos-core

http

Caddy-based web server providing web console hosting, file sharing, and health check endpoints. Use when configuring HTTP access, setting up file sharing, or troubleshooting web connectivity.

1,038 113
Explore
zylos-ai/zylos-core

health-check

System health check dispatched by the activity monitor via Control queue. Checks PM2 services, disk space, and memory usage. Use when receiving a control message containing "health-check".

1,038 113
Explore

Didn't find tool you were looking for?

Be as detailed as possible for better results