Bug Description
Today, Claude Code’s terminal UI is highly polished but largely fixed in layout. As a power user working in longer-running sessions, I often want persistent, glanceable context without interrupting the main interaction flow.
I would like to propose support for optional, extensible UI regions—for example bottom-left, bottom-right, or side panels—that Claude Code (or future plugins) could populate.
Potential uses include:
Live context summary (current task, assumptions, constraints)
Tool / file activity indicators (files read, edited, staged)
Session metadata (model, workspace, repo root, branch)
Progress indicators for multi-step agent tasks
User-defined notes or pinned instructions
Key characteristics that would make this especially useful:
Opt-in and minimal by default (no added clutter)
Dynamically hideable / toggleable via keybinds
Non-interruptive (no impact on main input/output stream)
Safe for terminal resizing and SSH usage
From a UX perspective, this would significantly improve observability and situational awareness during longer Claude Code sessions, especially in larger codebases or monorepos.
I believe this could be implemented incrementally (even starting with a single optional status panel) and would unlock future extensibility without compromising the current clean UI.
Environment Info
- Platform: darwin
- Terminal: Apple_Terminal
- Version: 2.1.7
- Feedback ID: b0dcef12-590f-485e-8d73-b5200c53b39d
Errors
Bug Description
Today, Claude Code’s terminal UI is highly polished but largely fixed in layout. As a power user working in longer-running sessions, I often want persistent, glanceable context without interrupting the main interaction flow.
I would like to propose support for optional, extensible UI regions—for example bottom-left, bottom-right, or side panels—that Claude Code (or future plugins) could populate.
Potential uses include:
Live context summary (current task, assumptions, constraints)
Tool / file activity indicators (files read, edited, staged)
Session metadata (model, workspace, repo root, branch)
Progress indicators for multi-step agent tasks
User-defined notes or pinned instructions
Key characteristics that would make this especially useful:
Opt-in and minimal by default (no added clutter)
Dynamically hideable / toggleable via keybinds
Non-interruptive (no impact on main input/output stream)
Safe for terminal resizing and SSH usage
From a UX perspective, this would significantly improve observability and situational awareness during longer Claude Code sessions, especially in larger codebases or monorepos.
I believe this could be implemented incrementally (even starting with a single optional status panel) and would unlock future extensibility without compromising the current clean UI.
Environment Info
Errors