Junshi (军师)

Also ask, but fill in yourself if they skip or forget:

• Target venues: Which conferences/journals matter most to you?
  • If they don't answer: infer from their research area using references/venues.md. Pick the 4-6 most prominent venues for their field and tell them what you chose. They can correct you.
• arXiv categories: Which arxiv categories are most relevant?
  • If they don't answer: infer from their field using references/venues.md. Tell them what you picked.
• Preliminary results: Any early experimental results, observations, or hypotheses — even informal, partial, or surprising ones. These are often the most valuable input. Numbers, patterns, things that worked or failed unexpectedly, conjectures not yet tested. Tell the user: "A single surprising observation can unlock better ideas than a finished paper." Users can add results anytime: "add a preliminary result: [result]"

Never block on missing answers. Make a confident default choice and say so — e.g., "I'll watch NeurIPS, ICML, ICLR, and CVPR for you — let me know if you'd like to add or swap any."

2. Read the researcher's papers

If no papers folder was provided or it doesn't exist, skip this step entirely and proceed to step 3 with an empty profile.

Use the Read tool (with page ranges for large files) or Bash with pdftotext on each PDF in the specified folder. For each paper, extract:

• Core technical contribution and methodology
• The frameworks and tools the researcher is fluent in
• Open problems or limitations explicitly or implicitly acknowledged
• What assumptions they rely on
• Research trajectory — what thread connects the papers?

3. Build the research profile

Note on paths: The skill is installed at ~/.claude/skills/research-junshi/ and all user data is stored under ~/.claude/research-junshi/.

Save to ~/.claude/research-junshi/profile.md:

# Research Profile

## Research Area
[Field(s) the researcher works in]

## Target Venues
[List of conferences/journals to monitor]

## arXiv Categories
[List of arxiv category codes, e.g. cs.CL, cs.LG]

## Research Themes
[Key topics and directions from the papers]

## Methods & Frameworks Used
[Mathematical/technical frameworks the researcher is fluent in]

## What's Already Been Done
[Brief list of contributions — to avoid redundancy]

## Open Problems in Their Work
[Gaps, limitations, and future work mentioned across papers]

## Research Taste
[What kinds of contributions do they value? Theory? Empirical? Applications?]

## Problem Statement
[The rough problem the user gave you, with your interpretation]

## Preliminary Results
[All results, observations, and hypotheses the user has shared.
Append new entries — never overwrite old ones.
Format:
- [Date] [Observation, as specific as possible]
  → [Your interpretation: what does this suggest or imply?]
]

## Last Updated
[Date]

4. Save config

Save to ~/.claude/research-junshi/config.md:

# Config
- Papers folder: [path]
- Problem: [problem statement]
- Research area: [field]
- arXiv categories: [comma-separated list]
- Target venues: [comma-separated list]

Daily Digest Workflow

On each daily run, load ~/.claude/research-junshi/profile.md and config.md first. Then:

Step 1: Search arXiv (last 24 hours)

Use WebFetch to query the arXiv API using the categories and keywords from the user's config.

Template URL (fill in categories and keywords from profile):