AI Engineering

MCP Builder helps you create Model Context Protocol servers that add tools, resources, and prompts for Claude. Use it when you want to connect Claude to external services or data sources.

Next-Best-Action designs event-driven recommendation engines for advisors. Use it to detect portfolio, life, market, and compliance triggers, rank competing actions, route them to the right person, and track acceptance.

Seedance 2.5 plans controllable video generation routes using Seedream storyboards, Seedance 2.0, and 2.5 when available. Use it for storyboard-to-video, reference-to-video, keyframed shots, extensions, and edits.

Universal Video Prompt writes one model-agnostic video prompt spec and compiles it for the target video model. Use it for cross-model prompt work, comparing models, or reusing one brief across providers.

Create Skill turns a skill idea into a complete repository skill with instructions, metadata, references, scripts, artwork, README wiring, and validation. Use it when creating a new skill or making an existing workflow reusable and consistent.

Lead turns an approved goal into coordinated agent work for autonomous builds. It plans, delegates implementation and review, and finishes with one pull request without writing or running code itself.

Lead Research orchestrates deep research by delegating scouting, verification, synthesis, and audit to subagents. Use it for brainstorming, technology choices, comparisons, and state-of-the-art surveys when you need a managed research workflow.

Map The Landscape builds a mental model of a topic or repository by outlining its boundaries, layers, actors, relationships, flows, and open questions. Use it when you need the big picture, need orientation, or want to understand how the pieces fit together.

Market Validation Audit runs an evidence-first market-fit and viability review for a project or idea and produces a decision doc. Use it to assess demand, competitors, timing, and next steps before spending more time building.

Paper Opportunity Radar runs a bounded literature sweep to audit paper claims, methods, integrity signals, and follow-on evidence. It is used to find credible, underexplored research ideas and separate them from hype, weak evidence, or contradicted results.

Planning interviews the user, explores the codebase, and turns unclear requests into implementation-ready plans. Use it for new features, refactors, migrations, or other non-trivial changes that need precise requirements and verification.

Repo System Map analyzes a repository’s implemented architecture and turns it into an interactive isometric system map with citations, dependency and data-flow paths, and a learner-focused explainer panel. It is used when you want to visualize or explain how a repo’s runtime, services, and deployment fit together.