Aggregate publications, student trajectories, funding, and lab culture into an HTML advisor report.
npx skills add https://github.com/jiadizhunine/deeptutor --skill deeptutor
Please help me install this Agent Skill. Skill: deeptutor Source: https://github.com/jiadizhunine/deeptutor/tree/main Command: npx skills add https://github.com/jiadizhunine/deeptutor --skill deeptutor
DeepTutor automates comprehensive evaluation of graduate advisors by aggregating publications, student trajectories, funding, and lab culture signals. It applies region-aware strategies (CN mainland vs international) and outputs a standalone HTML report generated from structured JSON, while supporting multiple agent ecosystems (Claude Code, Codex CLI, Cursor, OpenCode, and OpenClaw). It adheres to rigorous quality rules, sources every claim, and offers a full 10-phase workflow or a lite 6-phase option with a sharp critique and deal-breaker mechanism.
Prompt: "Investigate Prof. Smith at MIT Biology department" to generate a full advisor report.
High-intent search queries and answers about installing and using this skill.
To evaluate a potential PhD advisor, you need to analyze their publication records, student trajectories, funding, and lab culture signals. DeepTutor automates this comprehensive evaluation, aggregating the data to identify red or green flags for informed decision-making.
Yes, you can evaluate advisors at Chinese mainland institutions. The advisor evaluation process applies region-aware search strategies tailored for CN mainland versus international institutions to ensure accurate data aggregation and relevant lab culture analysis.
The best way to compare multiple graduate advisors is using a comparative mode that evaluates publication output and student trajectories side-by-side. This approach generates structured JSON rendered into a standalone HTML report, highlighting distinct deal-breakers and critiques for each advisor.
Yes, the advisor evaluation workflow supports multiple agent ecosystems including Claude Code, Codex CLI, Cursor, OpenCode, and OpenClaw. This compatibility allows you to run the 10-phase investigation or lite 6-phase option within your preferred coding environment.
To identify deal-breakers when analyzing lab culture and student outcomes, you apply a sharp critique mechanism to the aggregated publication and trajectory data. This process annotates specific red flags and critical issues to guide your advisor selection decision.
The difference between the full and lite advisor investigation workflows is the depth of evaluation. The full workflow uses a 10-phase investigation, while the lite option uses a 6-phase process, both outputting standalone HTML reports with sharp critiques and deal-breaker annotations.
bettercallfanMaps latitude/longitude trajectory data to Geohash grids for aggregation using Python's Geohash encoding.
LirsakuraAnalyze and synthesize intelligence on potential PhD supervisors from academic metrics and reviews.
renan-garciaMigrate AI agent configs into a unified .agents/ directory structure.
ionmidoriEvaluate LLM outputs using automated NLP metrics and human annotation frameworks.
