Skill UI

This skill provides comprehensive tools for planning studies by calculating the required sample size, minimum detectable effect (MDE), and statistical power. It supports both closed-form statistical tests (t-tests, ANOVA, chi-square) and advanced simulation-based designs (mixed models, logistic/Poisson regression). Use this when planning a study, submitting an IRB protocol, or justifying sample size for a grant.

A crucial post-hoc review skill designed for AI-generated Markdown prose. It audits content against a comprehensive set of 21 stylistic and structural rules using both deterministic mechanical checks and deep semantic judgment. Users can audit a single file to get a detailed scorecard, or compare two drafts (A/B) for regression testing. Furthermore, it can optionally generate a polished, revised copy in a separate file, ensuring the integrity of the original document while significantly improving linguistic quality and adherence to style guides. Ideal for refining AI-written documentation, ...

This skill enables the creation of a robust 'harness' system for coding agents within a repository. It addresses the need to prevent repeated AI-generated coding mistakes by implementing durable instructions, guardrails, regression tests, and failure memory. It treats the target repository as the source of truth, ensuring consistency and maintaining high reliability across all automated contributions.

This skill outlines a systematic, multi-phase workflow for diagnosing hard bugs and performance regressions in complex codebases. It emphasizes the critical necessity of building a tight, deterministic feedback loop (Phase 1) before moving to hypothesis generation. The process guides users through reproducing the failure, minimizing the scope of the issue, and forming actionable, ranked hypotheses.

A systematic, non-mutating investigation designed for complex software failures, including bugs, regressions, crashes, and flaky behavior. This skill launches four specialized, read-only sub-agents in parallel to perform deep root-cause analysis, trace code paths, pinpoint the smallest failing seam, and predict the quickest path to confirming or fixing the issue. It is designed purely for diagnosis and hypothesis generation.

Utilize this skill to assess the health, adoption, and performance metrics of published EAS Updates directly via the CLI. It reports critical data points including crash rates, unique user counts, install volumes, payload size, and the split between embedded and OTA users, making it essential for post-publish monitoring and regression detection.

This skill helps diagnose performance bottlenecks, slow jobs, and expensive queries across various data platforms (Airflow, dbt, Databricks). It uses a tiered investigation approach, starting with broad discovery to pinpoint problematic jobs or queries, bridging to affected tables, and finally drilling down to identify root causes, latency trends, and change-related regressions.

This skill performs a comprehensive, read-only, multi-agent review of a git diff or file scope. It launches four specialized sub-agents—covering Regression, Security, Performance, and Contracts—in parallel. The main agent then aggregates, filters, and summarizes critical findings, ensuring the code change is robust, secure, and maintainable. Ideal for pre-commit or Pull Request review workflows.

Run Lighthouse audits across every URL on a site with Unlighthouse, then aggregate performance, accessibility, best practices, and SEO scores. Ideal for local CI checks, offline CWV measurement, and post-deploy regression testing.

This skill generates comprehensive visual regression test configurations using SmartUI for accurate screenshot comparison on cloud platforms. It is framework-agnostic, supporting major tools like Playwright, Selenium, Cypress, and Puppeteer. Use it when detecting UI changes, ensuring visual consistency, or running automated visual tests across multiple viewports and browsers.

Diagnoses full-stack failure modes in agent and LLM applications, including wrapper regressions, memory contamination, tool discipline failures, hidden repair loops, and rendering corruption. Produces severity-ranked findings and code-first fixes for production agent systems.

This comprehensive guide outlines a rigorous, human-in-the-loop engineering workflow for leveraging advanced AI agents to complete complex implementation tasks. It mandates defining clear completion criteria, decomposing work into manageable units, and strategically routing models (Haiku, Sonnet, Opus) based on task complexity. The workflow emphasizes continuous evaluation, regression testing, security review of invariants and edge cases, and strict cost discipline to ensure robust and reliable software development using AI assistance.