Skill UI

This skill provides a scaffold for building complex, composable recommendation, ranking, and feed pipelines. It implements the robust six-stage pattern (Source→Hydrator→Filter→Scorer→Selector→SideEffect), which is the industry standard for determining "top K items." Use it for social feeds, content CMS ranking, RAG reranking, or any system needing complex candidate prioritization.

A comprehensive intelligence tool designed for clinical researchers to navigate the complex NIH grant landscape. It utilizes a detailed 6-step intake process (covering research idea, career stage, and data status) to perform a multi-faceted positioning analysis. The system maps research to appropriate NIH institutes using RePORTER data and generates an editable Word document containing actionable mechanism recommendations, submission timelines, and strategic guidance for funding applications.

A comprehensive tool designed for academic literature orientation. It searches papers using free keyless APIs (PubMed and OpenAlex) and builds a structured 'launching pad' guide, rather than a finished review. It helps researchers define their scope using frameworks like PICO/SPIDER, ensuring a confident and systematic starting point for deep research in unfamiliar fields.

A comprehensive tool for Bid Managers and Proposal Leads to structure responses to RFPs, RFIs, and RFQs. It systematically parses complex requirements, maps product proof-points to identify strategic gaps (GAP audit), develops win-themes using the Shipley method, and calculates a defensible winrate estimate to guide a decisive bid/no-bid decision. Focuses on strategy, not narrative writing.

A comprehensive tool for the full lifecycle management of applications within the Lark ecosystem (Feishu/Miaoda). It supports three development paths: local full-stack development, HTML static hosting, and cloud-based AI session generation. Use this when creating, updating, deploying, or publishing any complex system, tool, or platform application.

This advanced skill manages complex, large-scale development tasks by coordinating multiple, specialized subagents in parallel. It systematically decomposes the goal into isolated missions, manages resource quotas (token usage), and handles error recovery across multiple files and components. Ideal for large feature builds, system refactoring, or complex applications requiring multiple, concurrent roles (Planner, Builder, Tester).

This skill operates as a blunt, market-obsessed operating system, emulating the rigorous thought process of Marc Andreessen. It is designed to pressure-test ideas, ventures, and product strategies by strictly adhering to market-first principles (Market > Team > Product) and focusing solely on achieving Product/Market Fit (PMF). It rejects platitudes, refuses to validate premises, and is ideal for determining if a business concept has real market traction or if the venture is fundamentally flawed.

This skill conducts a comprehensive audit to identify and prioritize automation opportunities within a user's existing workflow. It guides users through a structured intake (covering roles, repetitive tasks, and pain points) and systematically scans connected tools (like Gmail, Slack, Calendar, and CRMs) to surface actionable, ranked quick wins, minimizing implementation effort and maximizing impact.

This skill guides the creation and maintenance of a `.geminiignore` file. By systematically filtering out machine-generated noise—such as build artifacts, dependency lock files, logs, and caches—it significantly optimizes the AI agent's context window. This process improves processing speed and, critically, reduces token consumption costs (FinOps).

This skill runs the Bumblebee scanner on macOS/Linux to perform read-only inventory and exposure checks. It detects compromised packages, developer tools, and host configurations against known security advisories. It is primarily used for supply-chain incident response and vulnerability assessment, providing developers with a clear picture of local system exposure.

A structured playbook for improving the time and space complexity (Big-O) of existing, slow code. This skill guides developers through systematic refactoring, focusing on identifying bottlenecks like nested loops or N+1 queries. It emphasizes applying transformations one at a time, maintaining test coverage, and measuring performance gains to ensure correct and efficient code.

This skill converts complex, narrative technical documents, research excerpts, or natural language problem descriptions into a highly structured Mathematical Problem Specification (MPS). It systematically extracts variables, constraints, objectives, and uncertainty parameters. By adhering to a strict Zero-Inference Protocol, it ensures all extracted elements are strictly grounded in the source text, making it invaluable for formalizing real-world optimization, data science, and engineering problems.