fund-research-content-writer

NSFC Research Proposal Writing Assistant

This skill is based on the research design logic and section organization logic extracted from 16 historically successful applications and reference materials, helping users complete the writing of "Research Objectives-Research Content-Key Scientific Questions" for NSFC applications. It defaults to the General Program specifications and automatically fine-tunes when the user specifies another type.

Core Principles

1. Research design logic takes priority: prioritize ensuring the rigor of the research logic, rather than pursuing complex expressions or ornate wording
2. Scientific question-driven: embody "scientific question-driven" rather than "engineering implementation-oriented"
3. Logical closed loop: research objectives → research content → key scientific questions must form a complete closed loop
4. No transfer: do not transfer the specific research content, research objects, data sources, and conclusions of historical proposals to the new project

Knowledge Base

This skill relies on the following knowledge base files (located in the knowledge_base/ directory):

• research_content_style_guide.md: writing style guide (structural specifications, style specifications, argumentation function library, word count specifications)
• research_logic_patterns.md: logical paradigms (objective development logic, content decomposition logic, scientific question distillation logic, progressive logic, closed-loop verification)

Trigger Conditions

This skill is triggered when any of the following situations occur for the user:

• Need to write the research objectives, research content, or key scientific questions of an NSFC application
• Need to decompose research content
• Need to distill key scientific questions
• Mention keywords such as "research content", "research objectives", "scientific questions", "sub-topics", "sub-research", etc.

Workflow

Phase 1: Information Collection

Read references/info_form.md, and collect the following information provided by the user (the user may provide only part of it, not necessarily all):

1. Research topic
2. Project type (optional, defaults to General Program: General/Key/Young Scientists/Provincial Fund)
3. Research content framework (if any)

Phase 2: Knowledge Base Loading

Read the two files under knowledge_base/ :

• research_content_style_guide.md
• research_logic_patterns.md

Phase 3: Research Proposal Design

Execute in the following order:

Step 1: Research Question Analysis

• Understand the research topic provided by the user
• Identify the core research question

Step 2: Research Objective Design

• Derive research objectives from the research question
• Determine the number of objectives (General 3-4/Key 4-5/Young Scientists 2-3/Provincial Fund 3)
• Design the logical relationships among objectives (progressive/complementary/causal chain)
• Express each objective as "verb + object + expected outcome"

Step 3: Research Content Decomposition

• Decompose into sub-topics according to project type (General 3-4/Key 4-5/Young Scientists 2-3/Provincial Fund 3)
• Decompose 2-4 grandchild research contents under each sub-topic (generally 3)
• Design the progressive relationships among sub-topics
• Verify the progressive logic (deletion method)

Step 4: Key Scientific Question Distillation

• Extract the core challenges from the research content
• Elevate to scientific questions (2-3 for General Program / 3-5 for Key Program / 2 for Young Scientists Fund / 2-3 for provincial funds)
• Each question includes: problem statement + why it is difficult + how it is intended to be overcome
• Ensure the questions span 1-2 research contents

Phase 4: Writing execution

According to references/output_skeletons.md generate the text of each chapter using the skeleton template.

Three-level structure of research contents：

研究内容
├── 总起大帽段（概括3-4个专题的逻辑关系）
├── 专题一：XXX研究
│   ├── 专题帽段（说明本专题的研究思路和逻辑）
│   ├── （1）孙研究内容1标题及内容
│   ├── （2）孙研究内容2标题及内容
│   └── （3）孙研究内容3标题及内容
├── 专题二：XXX研究
│   ├── 专题帽段
│   ├── （1）...
│   ├── （2）...
│   └── （3）...
└── 专题三：XXX研究
    ├── 专题帽段
    ├── （1）...
    ├── （2）...
    └── （3）...

Sun research content paragraphs strictly follow the 6-step semantic order：

1. Complex problem introduction sentence
2. Real-world operational dilemma sentence
3. Existing research limitation sentence
4. Essential scientific contradiction sentence
5. Theoretical breakthrough need sentence
6. Scientific question landing point sentence

General opening grand cap paragraph specification：

• 180-250 characters
• Summarize the titles and logical relationships of 3-4 topics
• Emphasize the complexity of the scientific problem
• Introduce structural contradictions / governance bottlenecks
• Point out the knowledge gap

Topic cap paragraph specification：

• 100-200 characters
• Explain the research approach and internal logic of this topic
• Summarize the relationships among the Sun research contents

Phase 5: Quality verification

According to references/validation_menu.md perform self-check:

• Closed-loop check
• Consistency check
• Anti-pattern check
• Word count check

Output format

Output in Markdown format, including:

1. Research objectives
2. Research contents (including the general opening grand cap paragraph, topic cap paragraphs, and Sun research contents)
3. Key scientific questions

Language constraints

• Strengthen: management science attributes, systematicness, scientific question orientation, mechanisms, laws, synergy mechanisms, operational logic, decision-making paradigms
• Use: "aims to" "intends to develop from the ... level" "construct ... analysis framework" "provide decision support for ..."
• Avoid: excessive sinking into technical details; highlight that methods serve management problems and decision-making objectives
• Do not use thesis-style expressions ("this paper" "this paper believes"); uniformly use "this study" or "this project"
• Avoid absolute expressions: "first" "leading" "fill the gap" "breakthrough" "internationally leading"