[FEATURE] Add inline autocomplete for /slash commands and other commands that uses `/`

Preflight Checklist

• [x] I have searched existing requests and this feature hasn't been requested yet
• [x] This is a single feature request (not multiple features)

Problem Statement

I constantly need to reference specific files, directories, MCPs, and execute slash commands while working with Claude Code. Currently, I have to remember the exact file paths
and command names, then type them out manually. This creates significant cognitive load and interrupts my workflow because I must:

1. Remember exact file paths and directory structures
2. Recall precise slash command names
3. Switch between my task and typing mechanics
4. Deal with typos and corrections that break my flow

This manual process is inefficient, error-prone, and breaks the natural conversation flow with Claude Code.

Proposed Solution

I'd like Claude Code to implement IDE-style inline autocomplete with these main triggers:

/ Symbol - Slash Command Discovery:

• When I type / in the prompt, show available slash commands
• Display command descriptions alongside names
• Allow human-readable language selection (e.g., "commit my changes" maps to /git:git-commit)
• Support command chaining in natural language
• Execute based on prompt context rather than rigid syntax

Interface Requirements:

• Dropdown appears immediately after typing @ or /
• Results update in real-time as I type
• Escape key closes the dropdown
• Tab/Enter to select items
• Minimal visual disruption to the prompt

Alternative Solutions

Currently I work around this by:

1. Opening separate file explorer to copy paths
2. Keeping a personal list of common slash commands
3. Using tab completion in terminal before copying to Claude Code
4. Making multiple attempts due to typos or forgotten names

These workarounds are time-consuming and break the conversational flow.

Priority

High - Significant impact on productivity

Feature Category

Interactive mode (TUI)

Use Case Example

_No response_

Additional Context

• Fuzzy search should be tolerant of typos and partial matches
• Commands should be discoverable based on intent, not just exact names
• The feature should work seamlessly with existing prompt functionality
• Performance should be responsive even with large codebases