learn-english skill for agents
A prompt-engineered framework for vocabulary memorization, sentence analysis, and collaborative corpus building.
Abstract
This project presents a prompt-engineered framework for agent-assisted English learning. The design objective is to improve vocabulary retention efficiency through rapid mnemonic generation, sentence-level interpretation, and bilingual explanation pipelines. The current implementation is a prompt protocol and has not yet been packaged as a formal agent skill.
Demonstrations
- Demo: Click Here
- Try More Advanced Prompts: Click Here
Community Contribution Protocol
If you have a GPT-4 Plus account, you are invited to contribute by uploading your vocabulary-list outputs to repository issues. This collaborative mechanism allows learners without GPT-4 Plus access to benefit from curated results.
System Modules
| Module | Function |
|---|---|
| Sentence Analyze (2024-04-06) | Structured analysis of sentence grammar and semantics |
| Chinese Version | Chinese instructional prompt set |
| English Version | English instructional prompt set |
| Result-EnglishVersion | Output archive for English-mode responses |
| Result-ChineseVersion | Output archive for Chinese-mode responses |
| Partial Result | Demonstration subset for quick inspection |
The demonstration output is designed for direct migration into spaced-repetition tools (e.g., Anki), where mnemonic variants can be reviewed as cards.
Prerequisites and Usage Boundary
- Minimum learner level: A1 English proficiency.
- Recommended model tier: GPT-4 Plus for best mnemonic relevance and linguistic accuracy.
- If GPT-4 Plus is unavailable, use precompiled outputs stored in
/GRE,/TOEFL, and related folders. - Avoid running GPT-4-specific prompts on lower-capability models when accuracy is critical.
Quick Start
I. What Is Mr.G?
Mr.G is a prompt series for English learning with three major functions:
- Automatic generation and switching of mnemonic strategies for hard vocabulary (targeting one-minute memorization cycles).
- Analysis of difficult vocabulary, sentences, and paragraphs.
- Repository-based restoration of mnemonic corpora for GRE, TOEFL, IELTS, and Duolingo-style preparation.
II. Operational Procedure
- Open the
/promptsdirectory and select the prompt file aligned with your task. - Input target vocabulary, phrasal verbs, sentences, or short paragraphs.
- Run the corresponding model variant:
- Use prompts labeled for GPT-3.5 when operating on GPT-3.5.
- Use
GPT3.5-APIversions when token conservation is required. - Use GPT-4 Plus prompts for full feature access.
- Export outputs to Anki, Notion, or Obsidian for long-term review and traceability.
III. Extended Function
-
G-GPT4.md: grouped-vocabulary processing workflow (updated).
Recommended Note-Taking Stack
- Anki
- Notion
- Obsidian
These tools support structured storage, iterative review, and contribution back to the shared corpus.