Undergraduate & master's degree thesis writing
When to use
- The user needs to review/optimize a thesis outline (distinguishing science and engineering / liberal arts)
- The user needs to imitate writing thesis chapters (introduction, abstract, experimental chapters, etc.;Liberal artsalso includes literature review, case analysis, countermeasures and suggestions, and liberal arts versions of the introduction/abstract)
- The user needs reference matching and GB/T 7714 format
- The user needs polishing, abbreviation, expansion, and de-AI-ification (use experimental polishing for experimental chapters and liberal arts polishing for liberal arts chapters)
- The user needs Chinese-English mutual translation
- The user needs to extract structured information from the thesis (for defense PPTs, etc.)
Step 1: Identify the task type
Select the corresponding reference file to execute according to the user's needs.
Outline review
Structural imitation writing
General(applicable to both science and engineering / liberal arts)
Science and engineering (science-*)
Liberal arts (liberal-*)
Problem and format check
| Task |
File |
| Problem and format check (typos / punctuation / layout / figure, table, and formula numbering / chapter references / duplicate paragraphs / terminology consistency) |
check-issues-and-format.md |
References and integration
Polishing
Abbreviation, expansion, anti-AIGC
Translation
Structured information extraction
| Task |
File |
| Extraction (for defense PPTs, etc.) |
extract-structured.md(after completion, you may ask whether to chain to defense PPT generation) |
Step 2: Collect input
- Multi-input task(For example, if imitating an introduction, 【reference sample essay】【paper outline】【personal draft】 are required): Parse the 【】 markers in the user message, ask follow-up questions about missing items, or read the workspace files specified by the user, and execute only after everything has been collected.
- Single-input task: Extract directly from the user message or pasted content.
Step 3: Execute and output
Read the Prompt in the corresponding reference, execute according to its Role/Task/Constraints/Output Format, and output the main text or structured result that meets the requirements.
Notes
- Word adaptation: Markdown is strictly prohibited; add spaces between Chinese and English/numbers/formulas; use full-width Chinese punctuation.
- De-AI-ification: Avoid "significantly improved", "greatly enhanced", "outstanding performance", "moreover", etc.
- Zero information loss: It is strictly prohibited to delete experimental parameters, data, or core arguments.
- Writing order: Do first, then write; from the inside out—first complete the methods chapter and experiments, then write the introduction and abstract.