Skill UI

This skill monitors production cost burn rate trends over time by binning session spend into configurable buckets. It calculates window-over-window deltas and triggers exit code 1 if the latest bucket exceeds the prior mean by a configurable percentage. Useful for CI integration to detect spending acceleration before budget limits are hit, distinct from benchmark drift tracking.

A comprehensive guide detailing best practices for ensuring the reproducibility of computational claims in Large Language Model (LLM) papers. It outlines a layered approach (Exact, Budgeted, Drifting) for verifying results, providing protocols for handling open-weight checkpoints, proprietary API models, and documenting compute costs and failure modes honestly.

Gate fine-tuned model checkpoints with data-quality checks, drift budgets, paired arena comparison, and forgetting remediation before promotion.