Computer Science Thesis Assistant

Problem

When writing computer-science coursework, materials are often scattered: project facts are hard to isolate, template style is inconsistent, chapter length drifts, reference ratios become unstable, figure and table numbering breaks, and the final draft stays in Markdown instead of a deliverable Word file.

How It Works

The skill makes the process explicit. First it locks input priorities, such as school templates, previous samples, user requirements, and default rules, and asks the user to resolve conflicts. It then reads project code and documentation to create a fixed fact base, so later chapters do not invent capabilities. It analyzes sample papers and templates beyond the table of contents, including fonts, sizes, line spacing, paragraph spacing, heading levels, caption styles, table content, and code block presentation. Before drafting, it returns a proposed outline, per-chapter word targets, and style decisions for confirmation. During writing, it maintains a chapter word table and compresses sections that exceed targets. It prioritizes required diagrams and tables, such as system architecture, E-R diagrams, key flows, data tables, and test cases; if mermaid or plantuml exists, it renders real images when possible and can capture real system screenshots when browser automation is available. References are constrained to post-2020 sources, usually 10–12 Chinese and 3–5 English entries, with verifiable items preferred. Final output targets .docx delivery and checks chapter completeness, word count, figure/table numbering, reference balance, and leftover placeholders.

Boundaries

It fits best when you have project materials, a template, or a sample paper, especially for school-format Word deliverables. Quality drops if project facts, citations, or style requirements are missing. It is not a plagiarism checker, anti-detection tool, or academic compliance auditor, and authors still need to verify references and claims.