run-cross-cluster-comparison

Community

Validate cross-cluster claims with regime matching and overlap analysis.

Authorj-mckerracher
Version1.0.0
Installs0

System Documentation

What problem does it solve?

Cross-cluster claims are only defensible when restricted to overlapping workload regimes. Follow this procedure exactly. Non-negotiable rule: No global cross-cluster claims without regime matching + overlap reporting. If you skip this and make a claim like "transfer from Anvil to Conte works," it will be rejected in review.

Core Features & Use Cases

  • Define a workload regime for each job using a shared taxonomy (cpu_standard, cpu_largemem, gpu_standard, gpu_largemem, unknown).
  • Compute overlap between source and target domains by training a domain classifier (e.g., logistic regression) on a curated feature set and applying a predefined propensity band to identify the overlap region.
  • Evaluate and report key metrics (AUC, target overlap coverage, KS statistics per feature) and clearly separate within-overlap performance from full-distribution results.

Quick Start

Run the cross-cluster comparison workflow on your dataset to assess regime matching and compute overlap.

Dependency Matrix

Required Modules

None required

Components

Standard package

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: run-cross-cluster-comparison
Download link: https://github.com/j-mckerracher/agent-research/archive/main.zip#run-cross-cluster-comparison

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.