What problem does it solve? Before writing a specification or implementation plan, developers often lack up-to-date documentation on the .NET APIs, LLM provider SDKs, and Azure services involved. This Skill gathers cited technical context from Microsoft Learn and Context7 so specs and plans are grounded in real documentation. ## Core Features & Use Cases - Topic Extraction: Derives search topics from user input or the current spec directory, focusing on .NET APIs, LLM provider APIs, protocols, Azure services, and NuGet packages. - Dual-Source Research: Queries Microsoft Learn (docs and code samples) and Context7 (library IDs and API references), collecting top results with URLs and key excerpts. - Structured Findings: Writes a research-docs.md file into the spec directory with sources, relevance notes, and a synthesis summary. - Fail-Soft Design: Continues gracefully when an MCP server is unreachable and never blocks the calling phase. - Use Case: Before specifying a new Cohere embedding feature, run this Skill to pull current Cohere SDK docs and .NET embedding patterns into specs/NNN-*/research-docs.md for the planner to cite. ## Quick Start Ask the assistant to research documentation for the current feature, for example: "Research documentation for adding Cohere image embedding support before we write the spec."
