What problem does it solve? Engineering teams often design or implement features without deeply understanding prior art, leading to spec drift, rework, and unverified assumptions. This Skill captures upstream knowledge rigorously by vendoring reference repositories and producing load-bearing research memos that specs and implementations can cite. ## Core Features & Use Cases - Repository Vendoring: Adds upstream repos as git submodules under ./vendors, pinned to specific commits for reproducible citation. - Three Memo Types: Produces spike memos (single time-boxed questions validated with runnable code), study memos (deep architectural dives with file:line citations), and survey memos (web/docs research with dated sources). - Structured Templates & Diagrams: Enforces memo templates with architecture maps, hot-path walkthroughs, ASCII diagrams, and explicit adopt/avoid decisions. - Use Case: Before designing a logging subsystem, ask the agent to study how tokio-rs/tracing resolves dispatchers; it vendors the repo, reads the hot path, and writes docs/research/study-tracing-dispatcher.md with cited evidence. ## Quick Start Use the research skill to vendor the tokio-rs/tracing repository and write a study memo on how it resolves the dispatcher per call site.
