Skill UI

This skill analyzes your data warehouse to pinpoint stale, unused, or redundant tables that contribute to unnecessary storage costs. It classifies waste patterns (e.g., Zombie, Temporary) and computes safety tiers, helping users understand which assets can safely be dropped or archived to optimize spending and improve data governance.

This skill wraps the `getCompactContext` functionality to efficiently retrieve a compacted ReasoningBank context, specifically optimized for cost-analysis queries. It generates a detailed report summarizing critical metrics such as tokens saved, memories retrieved, and agentic-flow availability. It is essential for monitoring and optimizing resource usage in complex AI workflows, providing deep visibility into the efficiency and cost of the context retrieval mechanism.

Lists every session in cost-tracking with start time, message count, top model, and total cost. Use it for per-conversation spend analysis, retrospectives, and project rollups.

This skill performs multi-baseline counterfactual cost analysis for AI model routing. It compares actual session spending against hypothetical scenarios using single models like Haiku, Sonnet, or Opus. It calculates savings percentages to determine if routing decisions are cost-effective. Negative savings indicate over-escalation, while positive savings show router efficiency. Useful for quarterly reviews, CI gates, and validating routing configurations.

This skill analyzes benchmark run data stored in JSON files to track performance drift over time. It calculates deltas between the first and last runs for metrics like win rate, latency, escalation rate, and LLM-baseline cost. The tool helps identify regressions before release, after corpus expansion, or following agent upgrades by flagging significant changes in speedup or latency compared to baselines.

This tool performs deep, multi-touch attribution analysis to calculate comprehensive Return on Investment (ROI), Return on Ad Spend (ROAS), and Customer Acquisition Cost (CAC) across all marketing channels. It supports various attribution models (e.g., linear, time-decay) and projects long-term value using LTV, providing data for budget optimization, executive reporting, and identifying efficiency gaps.

A comprehensive framework for translating reduced-form causal estimates (e.g., elasticity, treatment effects) into explicit policy statements. It guides researchers on building clear welfare, cost-benefit, or optimal-policy models, defining sufficient statistics, and articulating distributional effects, ensuring policy claims are rigorously grounded in empirical data.

A guide for advanced data analysis in policy evaluation, extending beyond standard causal estimates. It emphasizes rigorous reporting of cost-benefit ratios, distributional consequences (who gains and who pays), robust heterogeneity testing, and honest uncertainty quantification necessary for high-level policy academic journals. Essential for turning effect sizes into decision-relevant policy narratives.

A comprehensive guide designed to help researchers determine the optimal top-tier conference venue for their computer architecture work. It provides detailed criteria—such as mechanism-residency tests, hardware cost analysis, and scope boundaries—to guide authors in correctly placing their research between venues like MICRO, ISCA, ASPLOS, and DAC.

A rigorous decision-making methodology that quantifies the trade-offs when allocating scarce resources (time, budget, attention). Instead of merely evaluating a chosen path, this framework forces the user to identify the single best alternative forgone, calculating the net value delta (chosen value minus best forgone value). It is essential for feature prioritization, major commitments, and strategic planning to prevent decisions based solely on immediate perceived gains.

This framework guides decision-makers to structure their thinking by classifying decisions based on the cost and feasibility of reversal. It emphasizes treating low-cost, reversible choices (two-way doors) quickly, while reserving deep deliberation for high-cost, irreversible commitments (one-way doors). It helps mitigate analysis paralysis and premature over-engineering.

This methodology allows engineers and analysts to conduct controlled counterfactual simulations when real-world testing (such as catastrophic failures, extreme load, or irreversible changes) is too rare, costly, or impossible. By isolating a single variable and tracing the mechanistic chain step-by-step, it helps surface critical system breakpoints, invariants, and necessary controls before deployment.