Ever wonder how much Claude API costs you're racking up in a session, or whether TLDR is actually saving you money? This drops a detailed dashboard right in your chat showing real dollar amounts spent, token usage breakdown, and concrete savings from compression. You get file read comparisons (like 1.3M tokens crushed down to 20.9K), cache hit rates, and sparkline history. Most useful when you're trying to optimize costs or debug whether TLDR hooks are firing. The output is surprisingly readable with progress bars and emojis marking which model you're on. If you're running long sessions with big codebases, the savings numbers might surprise you.
npx skills add https://github.com/parcadei/continuous-claude-v3 --skill tldr-stats