Paper Context Resolver

When you're trying to reproduce a deep learning paper and the README doesn't quite tell you which dataset split they used or how they mapped their checkpoints, this skill fills that specific gap by pulling answers from the paper itself. It's designed as a helper that gets called only when you hit a concrete reproduction blocker, not for general paper summaries or to second-guess what the repo already documents. The useful bit is that it explicitly flags conflicts between what the paper says and what the README claims, which happens more often than it should. Think of it as a surgical reference lookup rather than a research assistant.

Install to Claude Code

npx -y skills add lllllllama/rigorpilot-skills --skill paper-context-resolver --agent claude-code