I built a zero-token watcher that shows whether your Claude sessions are actually working — every subagent, its runtime, and its token spend. No hooks, no server. MIT.
Why CEREBRO kept it
Token spend monitoring for Claude subagents
The text below is an automated extraction of the article at https://www.reddit.com/r/ClaudeAI/comments/1v29lkx/i_built_a_zerotoken_watcher_that_shows_whether/, stored verbatim in the public cerebro-vault repository. Copyright remains with the original publisher (reddit.com).
I built a zero-token watcher that shows whether your Claude sessions are actually working — every subagent, its runtime, and its token spend. No hooks, no server. MIT.
Backlinks
Appeared in 1 briefing