Skill UI

This skill generates a detailed, multi-phase plan for complex, multi-file refactoring projects. It ensures safety by first analyzing the codebase, identifying dependencies, and structuring changes in a safe sequence (e.g., Types first, then implementations, followed by tests and cleanup). It forces the AI to output a comprehensive plan, including rollback steps and verification points, and waits for user confirmation before making any actual code changes.

This advanced agent acts as a highly constrained implementation expert, designed to execute complex, multi-step plans with absolute fidelity. It strictly adheres to the provided workflow, making only specified changes, updating the plan status in real-time, and running tests until completion. Ideal for disciplined software development, detailed project execution, or formal process adherence.

Process videos, audio, images, and documents on Transloadit’s cloud using 86+ robots to encode formats, generate thumbnails, resize/watermark, run OCR, speech-to-text, moderation, and chain multi-step pipelines for scalable media automation.

A comprehensive skill designed to convert vague ideas, initial concepts, or high-level requirements into structured, detailed, and approved technical specifications. It enforces a rigorous, multi-step workflow, guiding users through context exploration, proposing multiple architectural trade-offs, and iteratively refining the design until a complete specification document is ready for the implementation phase.

A comprehensive guide detailing procedures for upgrading Abridge API versions and migrating between disparate Electronic Health Record (EHR) systems. It covers zero-downtime API versioning strategies (e.g., canary rollout) and outlines high-risk, multi-step procedures for ensuring seamless clinical documentation data continuity during EHR cutovers.

This guide provides step-by-step instructions for installing the Alchemy SDK and setting up secure API key authentication. It details how to connect to multiple major blockchain networks (including Ethereum, Polygon, Arbitrum, etc.) using Alchemy's robust infrastructure. Essential for developers building decentralized applications (dApps) that require scalable, multi-chain RPC access.

This comprehensive guide teaches developers how to interact with Claude AI using the official Anthropic SDK. It provides step-by-step instructions on basic message creation, implementing system prompts, and managing multi-turn conversation flows via the Messages API. Ideal for setting up a new Claude integration or testing core API functionality.

This guide details how to use sophisticated coding agents (such as Codex, Claude Code, or Pi) through a structured bash workflow. It provides methods for complex development tasks, including building new features, refactoring large codebases, and reviewing pull requests (PRs). Key best practices covered include managing pseudo-terminals (PTY), defining focused working directories (`workdir`), and executing tasks in the background for robust, multi-step automation.

This comprehensive guide details the setup and deployment workflow for Glean enterprise search across development, staging, and production environments. It emphasizes strict environment isolation by managing separate datasources, API tokens, and configurations for each stage. The skill covers best practices for promotion workflows, validation steps, and error handling, ensuring search relevance and data integrity during scaling and updates.

This guide provides advanced patterns for composing robust, multi-step Large Language Model (LLM) workflows using LangChain 1.0 core runnables. Learn to implement parallel retrievals (RunnableParallel), conditional routing (RunnableBranch), and state management. It also addresses common pitfalls like silent KeyErrors due to mismatched dictionary shapes, providing solutions using debug probes and type annotations for reliable RAG, classification, and extraction pipelines.

This guide teaches how to build robust, production-grade ReAct agents using LangGraph 1.0's `create_react_agent`. It specifically addresses common agent failure modes, such as preventing infinite loops (recursion caps) and ensuring proper error propagation. Learn to define typed tools using Pydantic, manage state with checkpointers, and set precise recursion limits. Perfect for users migrating from legacy AgentExecutor or building reliable, multi-step tool-calling workflows.

This comprehensive tool assesses the legal patentability of an invention by checking its novelty and non-obviousness against global prior art. It simulates professional patent examiner procedures, conducting detailed anticipation (novelty) and inventive step (non-obviousness) analyses using structured single-reference and combination tests. The output provides multi-jurisdictional assessments (US, CN, EP) and recommended claim amendments.