Run thesis research and data analysis with Claude Code skills

Commenters use Claude Code with custom skills for thesis work, plus agents that write experimental reports and extract, analyze and visualize large datasets.

Done withCCClaude Code

What they did

Commenters describe using Claude Code with project-level instructions in CLAUDE.md and custom skill files so the agent keeps research context, LaTeX conventions, and task-specific workflows across sessions. In practice they use separate agents or sessions to analyze papers, summarize and critique them against their own research, work locally on Overleaf TeX files, and build writing, extraction, analysis, and visualization agents for thesis reports and large datasets; the human sets up the skills and prompts and reviews the output.

How it went

The thread reports these workflows are useful for thesis writing, previous-work review, experimental reports, and massive-data analysis, but it gives no metrics or head-to-head results. One commenter warns research use needs explicit assumptions because the model can fill gaps confidently.

Worth knowing

The clearest reusable setup detail is to keep a CLAUDE.md at the project root and add skill.md files, since commenters say that preserves context across sessions and cuts cleanup.

Source: Reddit · Undated