Skill UI

An internal, non-user-invocable tool designed to validate candidate meta-skills. It runs critical G3 (positive) and G4 (negative) smoke tests by simulating meta-resolution using LLM-generated fixtures. This ensures the structural integrity and robustness of the skill in a cross-vendor environment.

This skill captures a fully self-contained static HTML snapshot from any running web application (e.g., React, Vue, Next.js). Using Puppeteer, it renders the live DOM, automatically inlining all CSS styles and converting images, assets, and backgrounds to base64 data URIs. This ensures the generated HTML file is completely self-sufficient and requires no external network connection, making it perfect for sharing UI states, creating documentation, or preparing assets for offline viewing.

This utility provides an extension point to inject custom query parameters into all live preview URLs generated by the Website Builder. Developers can use this to automatically append necessary data, such as signed tokens, tenant identifiers, or feature flags, ensuring the preview environment is fully configured for specific use cases. It supports complex async operations like fetching tokens from remote APIs.

Unship is a dedicated local workflow for front-end developers to compare multiple AI-generated UI alternatives (such as layouts, copy, or flows) directly within a running application environment. Instead of committing to one generated version, it adds temporary source-level variants, allowing for visual comparison. After the user selects the final version, the skill automatically cleans up all unused temporary code and attributes, ensuring a clean, production-ready codebase. Note: This tool is for local iteration only, not for production A/B testing or analytics.

This guide helps developers audit and clean up Vite + React projects initialized using the Lovable scaffolding tool. It systematically removes all boilerplate artifacts, including placeholder assets, dedicated dependencies (like lovable-tagger), generated documentation, and unnecessary components (e.g., unused Radix deps). The goal is to deliver a clean, fully owned, and production-ready codebase.

This skill guides the process of routing a complex goal through an AI's scored-memory loop using the `acc_act(runtime="solve")` function. It provides a structured workflow for deliberation, handling intermediate `brain_frame` results, and submitting continuous proposals. It emphasizes systematic reasoning and requires careful resolution of all generated frames to ensure a robust and verifiable solution.

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.

A specialized quality gate skill designed to review complex, multi-source outputs generated by the CrossFrame framework. It rigorously checks reasoning fidelity, evidence boundaries, source anchors, and structural continuity. Use this when advanced verification of logical rigor and source integrity is required, ensuring outputs meet strict internal protocols and avoid common structural failures.

This skill provides a rigorous, multi-layered audit for User Interfaces, proving layout balance and usability beyond subjective judgment. It combines a deterministic spatial audit (calculating centroid, optical-center, and pixel-oracle balance via explicit math) with a professional usability review based on Nielsen's 10 heuristics, conducted by a separate 'fresh-eyes' judge. Ideal for catching misaligned or poorly structured AI-generated designs.

This skill enforces a high standard for all generated code, prioritizing correctness, clarity, and idiomatic patterns across all languages. It helps minimize code bloat by removing unnecessary abstractions and verbose patterns, while also optimizing the agent's output by using targeted diffs and eliminating excessive conversational overhead. Use this for all code generation, refactoring, or review tasks.

A specialized tool designed to strip out predictable, statistical patterns common in AI-generated writing. It systematically rewrites overly polished or uniform text—such as blog posts, emails, or social media captions—by injecting genuine human signals, ensuring the final output reads like it was written by a real person.

This comprehensive guide details how to configure, integrate, and troubleshoot live email sending using Mailtrap. It covers two primary methods: the Email API (REST) and SMTP. Learn the differences between core email streams—transactional (app-generated notifications), bulk (promotional marketing), and batch requests (high-volume efficiency)—along with recommended integration patterns and specific technical endpoints.