NewKnow Teacher Assistant

The Administrative Burden Faced by Teachers

K12 educators in China often spend a significant portion of their time on repetitive, format-driven administrative and instructional planning tasks. This includes drafting standardized lesson plans for each class, creating exam papers with various question types for different assessments, analyzing student grade data to generate class reports, and writing individualized semester comments and parent communications. These activities consume valuable time that could otherwise be dedicated to direct teaching and personalized student guidance, and often involve manual conversion between document formats (Word/PDF).

A Structured Solution: NewKnow Teacher Assistant

This skill functions as a highly customizable AI agent designed to automate these workflows. Its core mechanics are as follows:

• Intent-Driven and Modular Capabilities: The skill defines 10 core capability modules (e.g., Lesson Plan Generation, Smart Question Setting, Academic Analysis), each activated by specific trigger keywords (e.g., "备课", "组卷"). Users interact through natural language (e.g., "Generate a Chinese lesson plan for 'The Enriched Xisha Islands'"). The system routes the request to the appropriate module.
• Structured Input and Templated Output: For practical and standardized outputs, each capability requires key inputs from the user (subject, grade, textbook version, specific topic). The system then generates content based on embedded Markdown templates (e.g., a lesson plan with objectives, process, and board design). This content can be automatically converted into formatted Word documents (.docx) using the built-in scripts/md_to_docx.py script, adhering to standards like A4 paper and fixed line spacing. Some capabilities (e.g., knowledge cards) also support PDF output.
• Embedded Professional Knowledge Base: More than a simple content generator, the skill embeds current educational policies (e.g., "double reduction" regulations), subject-specific exam design standards, student assessment criteria, and parent communication best practices. This ensures generated exams, comments, and communications align with contemporary educational requirements and scenarios.

Example Workflow and Important Boundaries

Typical Workflow (e.g., "Smart Question Setting"):
1. User Input: "Create a unit test for Grade 7 Math, Chapter 3 (PEP edition), including multiple-choice, fill-in-the-blank, and problem-solving questions. Medium difficulty."
2. AI Processing: Identifies the trigger "出题" and routes to Capability 2. Parses the subject, grade, version, chapter, question types, and difficulty.
3. Content Generation: Based on embedded "Subject-Specific Question Types and Design Standards," it generates a complete test with stems, options (for multiple-choice), and detailed solutions with steps (for STEM subjects).
4. Format Conversion: Automatically invokes scripts to generate a Word file named 试卷_第三章单元测试_20250513.docx.

Crucial Considerations:
* Teacher Review is Mandatory: AI-generated content, especially student evaluations and grade analyses, must be reviewed and adjusted by the teacher based on the actual classroom context before use.
* Textbook Version Dependency: The skill defaults to China's national unified textbooks (e.g., PEP for Math, Ministry of Education edition for Chinese). If using other versions, the user must specify this explicitly in the input.
* Privacy and Compliance: Generated documents are stored locally, but teachers must be mindful of data anonymization when involving student personal information. Outputs default to policies like "double reduction" (e.g., controlling homework volume, avoiding rankings).
* Toolchain Integrity: Document generation relies on scripts in the skill's scripts/ directory (e.g., knowledge_card_generator.py, paper_reviewer.py). Ensuring the skill environment is complete is essential.