@guanyangwangchecks…add-junshi-skill → master3 files · +424 −0review: @accesspath27-lgtmupdated 1w agoThis PR adds Junshi, a Claude Code skill for personalized research planning.
What problem it solves: Researchers often drown in new papers and summaries, but still struggle to identify the most promising next idea and first experiment.
Who uses this workflow: Researchers, PhD students, and technical teams who work in literature-heavy settings and want recommendations grounded in their own prior work.
Attribution / inspiration source: Inspired by the idea of an AI “junshi” (strategist): not replacing the researcher, but helping them think more clearly and act faster.
Example of how it’s used: Junshi reads a user’s local paper folder, builds a research profile, scans relevant literature, and produces a daily digest with 3–5 ranked ideas, first experiments, and main risks.
Placed under Data & Analysis to match the current README category structure.
Thanks! I added the author attribution to the README entry and included a concrete input/output example in SKILL.md.