Hash-7777/Hash-Medical-Reasearch-Agent-Skills ? reverse-engineered prompt

Build me a GitHub repo of drop in agent skills for medical research that helps an AI treat papers like a careful reviewer, not a chatbot.

I want plain text SKILL.md files that work with any agent or model, including Claude Code, Cursor, hosted models, and local Ollama. The repo should cover the full evidence workflow, from planning a search, running a reproducible PRISMA style literature search, extracting claims from sources, checking evidence coverage, spotting conflicts, grading every citation against its source, pooling studies only when they are actually comparable, and ranking source reliability. It should also include a skill for defending a RAG pipeline against prompt injection.

Please make the README clear and polished, with quickstart instructions, a simple explanation of why this matters, a list of the skills, and examples of how to use them. Include a small examples area and useful docs for integrations and FAQs. Keep the tone practical and honest, and if you need current details for any tool or integration, look up the latest docs online.