What problem does it solve?
Teams running AI workloads often discover budget overruns only after the money is spent. This Skill answers "when will we cross the line?" by computing a USD-per-day burn rate from recent session records and projecting spend across future horizons, so finance and SRE teams can act before budgets are exhausted.
Core Features & Use Cases
- Burn-rate computation: Reads session records from the cost-tracking namespace, filters to a configurable measurement window (default 7 days), and derives a per-day spend rate.
- Multi-horizon projection: Linearly extrapolates spend to 7d/30d/90d/365d horizons and reports days until 75%, 90%, and 100% of a configured budget is consumed.
- CI and dashboard integration: JSON output mode lets you gate builds (e.g., fail when full exhaustion is under 7 days away) or feed a budget dashboard for quarterly planning.
- Use Case: After shipping a large feature, re-run the projection to verify the daily spend rate has not accelerated beyond expectations, and alert if the 100% budget threshold is less than a week away.
Quick Start
Ask the AI to project my API spend for the next 30 and 90 days based on the last 7 days of usage and tell me when my configured budget will run out.