Feature Request: Make Claude Code commands available after they are created within the Claude Code session

Hey, I've run into this as well! It's a great point and would be a nice quality-of-life improvement.

From what I can tell from the docs and my own usage, Claude Code seems to scan for custom commands in the ~/.claude/commands/ and ./.claude/commands/ directories only when a session starts. It doesn't seem to have a file-watcher or a way to refresh its command list mid-session.

The current workaround is, as you've noted, to exit the session (Ctrl+D or typing exit) and start a new one. After you do that, the newly created command is available immediately.

It's interesting because some other configurations, like changes to hooks in settings.json, have a mechanism to be reloaded. The docs on hooks mention:

"Claude Code... Warns if hooks are modified externally [and] Requires review in /hooks menu for changes to apply"

It would be awesome if there were a similar mechanism for slash commands, maybe a /reload-commands command, or even better, if it automatically re-scanned the directory after using the Write tool to create a new command file.

Definitely a +1 from me on this feature request. Being able to chain commands agentically like "create a command to lint my files, then run that new command" would be a fantastic workflow enhancement.

Hey, I've run into this as well! It's a great point and would be a nice quality-of-life improvement. From what I can tell from the docs and my own usage, Claude Code seems to scan for custom commands in the ~/.claude/commands/ and ./.claude/commands/ directories only when a session starts. It doesn't seem to have a file-watcher or a way to refresh its command list mid-session. The current workaround is, as you've noted, to exit the session (Ctrl+D or typing exit) and start a new one. After you do that, the newly created command is available immediately. It's interesting because some other configurations, like changes to hooks in settings.json, have a mechanism to be reloaded. The docs on hooks mention: > "Claude Code... Warns if hooks are modified externally [and] Requires review in /hooks menu for changes to apply" It would be awesome if there were a similar mechanism for slash commands, maybe a /reload-commands command, or even better, if it automatically re-scanned the directory after using the Write tool to create a new command file. Definitely a +1 from me on this feature request. Being able to chain commands agentically like "create a command to lint my files, then run that new command" would be a fantastic workflow enhancement.

Its actually only agents and hooks that can reaload but its required that you go into the /agents or the /hooks menu and exit then it reloads - also settings is just added with 1.0.90 so that hot reloads - but yes it really would be nice to see that addition on commands and agents as well - MCP is different because it require a reconnect and there is not really a way for the mcp to restart even if there is code update and a MCP server can be so many different elements - so that is just way more complicated so a simple /mcp reconnect <optional mcp server> would be the best way to solve it on MCP server development

This issue has been inactive for 30 days. If the issue is still occurring, please comment to let us know. Otherwise, this issue will be automatically closed in 30 days for housekeeping purposes.

bump to stop the auto closer. I would really like this <3

Got my vote too, if we create or edit a Skill - Claude should be able to use it without a reload, or we should be able to do a in-session refresh.

VSCode Workaround (No Context Loss)
While true hot-reload within an active session isn't available yet, there's a quick workaround for VSCode users: Ctrl+Shift+P → Developer: Reload Window

This restarts all active Claude Code sessions in your workspace, making any newly created slash commands, rules, or skills immediately available.
The best part: Your conversation context is preserved.

The session resumes exactly where you left off - no need to re-explain what you were working on. Much faster than manually closing and reopening sessions.

This issue has been automatically closed due to 60 days of inactivity. If you're still experiencing this issue, please open a new issue with updated information.

This issue was closed incorrectly despite recent human comments. This behavior of the bot is reported at https://github.com/anthropics/claude-code/issues/16497. Please upvote that issue, so maybe it gets noticed.

This issue was closed incorrectly despite recent human comments. This behavior of the bot is reported at #16497. Please upvote that issue, so maybe it gets noticed.

Use a skill instead that is available as / command as well - and is way better than slash commands - and they hotreload on update and add

This issue has been automatically locked since it was closed and has not had any activity for 7 days. If you're experiencing a similar issue, please file a new issue and reference this one if it's relevant.