[BUG] Tool usage output no longer shows executed command in versions > 1.0.123

Yes, agreed. I'm also seeing this...

<img width="1197" height="295" alt="Image" src="https://github.com/user-attachments/assets/ae8dd329-9081-43a2-9d53-0e97f8e77d35" />

Flying blind, in auto mode I have no idea whats going on... Just hope for the best... I tried to enable verbose mode but id doesn't help.

Same here. Only seeing assistant outputs, tool calls are not visible. Only bash calls (that takes long, like running tests), only visible while the tool call happens, then disappears. No other tool calls are visible.

Persistent in 2.0.0

Same issue - very unfortunate. Maybe a concious choice to be more like Codex?

Its an absolute nightmare, I either manually have to approve each change, or blindly trust every intermediate step and hope for the best... Really I want to be able to watch the output and when its starts going off the rails stop and interject... or even just type a message for it to read at the next suitable time if its not super important... but we lost all of that oversight... its frankly scary now - you have no idea what its doing - Not sure how this is safe - sure you can go back but imagine it makes some tool calls and goes down a path that causes harm. You have not trace of what is going on so human can step in - that option is completely removed in auto mode.

Need to make it optional to not roll up tool calls into latest tool call which is what is happening now.

Kindly jump on this thread on x and make some noise if this kills your productivity - https://x.com/nkeilar/status/1973267989337444744

This bug is getting even more dumb and annoying. It's going on opposite direction as the new patch versions goes out.

Now aside of I can't see my actual tool call details, I see all the hooks about the tool calls which tells me nothing, it actually makes the Claude code output even harder to read and follow.

Claude code ppls, please fix this frigging issue.

!Image

Workaround for Hook Output Cluttering the Display

@mfyz and others - I found a solution to suppress the hook output that's been cluttering the display.

The hooks can be silenced in two ways:

1. Add suppressOutput flag - Set "suppressOutput": true in each hook configuration
2. Shell-level redirect - Append >/dev/null 2>&1 to the end of each command

Here's an example of what your hooks should look like in ~/.claude/settings.json:

  "PostToolUse": [
    {
      "matcher": "*",
      "hooks": [
        {
          "type": "command",
          "suppressOutput": true,
          "command": "/path/to/your/hook.sh >/dev/null 2>&1"
        }
      ]
    }
  ]

To update your settings:

Apply this pattern to all hooks in your settings file:

• PreToolUse
• PostToolUse
• Notification
• Stop
• SubagentStop
• PreCompact
• UserPromptSubmit

The double suppression (both the flag and shell redirect) ensures maximum output suppression and gives you a cleaner CLI experience.

This doesn't solve the root issue of tool commands not being visible, but it at least removes the hook noise that's making the output even harder to follow.

This appears to be resolved in 2.0.19

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.

Well it's an open UX question

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 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.