Claude Code cannot report its own token/quota consumption

Feature Request

Claude Code has no ability to monitor or report its own token or quota consumption mid-conversation.

Problem

For developers using Claude Code as a daily coding tool, there is no way to know when you are approaching a usage limit without leaving the tool and checking the Claude.ai dashboard manually. Claude Code itself cannot surface this information — it has no visibility into its own quota state.

Suggested Improvement

Claude Code should support two configurable monitoring modes — the developer must explicitly choose:

Pull (on demand): The developer asks Claude Code "how many tokens have I used?" and gets a current usage report.

Push (proactive): Claude Code warns the developer when approaching a configurable threshold (e.g. 80% of quota).

In both cases, if continued use would incur additional API charges beyond the subscription, the developer must expressly authorize those charges before Claude Code proceeds. Claude Code should never silently exceed a quota or trigger unexpected billing.

Why This Matters

Developers paying for a Claude.ai subscription to use Claude Code need operational visibility into consumption. Hitting an unexpected limit mid-session interrupts work with no warning — and unexpected charges are unacceptable without explicit consent.

Reported via Claude Code by a user who asked their Claude Code agent to file this gap directly — and noted that the agent lacked the capability to do so autonomously, which is itself a gap worth addressing.