UIST Experiments & Evaluation Auditor

•
1.1k
•

Designs and audits experimental evaluations for ACM UIST papers to align evidence with system claims and avoid flawed studies.

UIST Experiments is a specialized Claude Code skill tailored for human-computer interaction (HCI) researchers and systems engineers targeting ACM UIST and similar premier venues. It provides systematic guidelines for selecting the right evaluation methodology—ranging from technical characterizations and controlled within-subjects studies to developer walkthroughs and ablation suites—ensuring empirical evidence directly supports the core contribution. By conducting pre-submission evidence audits and identifying common reviewer objections, this skill helps authors replace uninformative ritual studies with robust, honest benchmarks across realistic operating envelopes.

Key Features

01Guides technical characterization across operating envelopes, failure boundaries, and tail distributions
02Pre-empts common reviewer objections regarding benchmark strawmen, task designs, and sample sizes
03Matches system claims (hardware, toolkits, pipelines, interaction techniques) to appropriate evaluation methodologies
04Audits pre-submission evidence to identify orphaned claims, underpowered studies, and review vulnerabilities
05Structures honest baseline comparisons and component ablation studies for interactive systems
061,090 GitHub stars

Use Cases

01Auditing an upcoming UIST or ACM CHI submission to ensure empirical claims have matching evidence
02Designing technical benchmark suites and ablation experiments for novel interaction pipelines or toolkits
03Diagnosing reviewer feedback during rebuttal preparation to address baseline and metric skepticism