[FEATURE]: /context — detailed token/cost breakdown panel per session component #41908

Requested feature

Add a /context slash command (inspired by Claude Code's /context) exposing a detailed breakdown of how the context window is being used: prompt, tools, and message history split by token counts, plus estimated cost.

Why

Today the sidebar (packages/tui/src/feature-plugins/sidebar/context.tsx) only shows the aggregate token count and cost for the latest assistant message (last.tokens.input + output + reasoning + cache.read + cache.write, plus money.format(cost())). There is no per-component visibility, so users can't tell whether prompt/system instructions, tools, or history is consuming the window.

Proposed behavior

• /context opens a dialog/sidebar breakdown: system prompt, tool definitions, session messages, and cached tokens each as a separate row with token counts and percentages of the model's context limit.
• Optionally surface estimated cost per component.
• Reuse the model context-limit metadata already read in context.tsx.

Benefits

• Diagnose context overflow before /compact is needed.
• Clear view of cost drivers per conversation.