Monitors real-time AI spending and projects total workflow costs based on per-phase performance data.
This skill provides essential financial visibility into AI-driven coding tasks by analyzing per-phase averages from completed work to estimate total project costs. It automatically tracks spending, projects remaining expenses based on workflow complexity, and displays a live budget status in the Claude Code HUD. By integrating customizable budget ceilings and smart profile suggestions, it empowers developers to maintain strict control over LLM API expenditures and switch to more affordable model profiles when costs trend high.
主な機能
013,216 GitHub stars
02Customizable budget ceiling warnings with overrun alerts
03Smart profile suggestions to switch to cost-optimized models
04Step-based averaging for high-accuracy estimation across different workflow types
05Real-time cost projection based on historical phase metrics
06HUD statusline integration for continuous spending visibility
ユースケース
01Identifying cost-heavy development phases to optimize model routing
02Managing API expenditures during high-token agentic coding tasks
03Setting strict financial guardrails for automated research and debugging sessions
