Cram Engine skill
A skill for exam cramming in one or two weeks: it builds a topic tree, teaches, tests and patches gaps.
- 180 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The skill only drives a conversation; it runs no code and calls no external services.
Install
- Works in
Manual install
npx skills add https://github.com/liuliu667/cram-engine --skill cram-engineSend the command to the agent in chat or run it in a terminal.
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skill only drives a conversation; it runs no code and calls no external services.
This is third-party code. Review the repository files before installing.
What it does
Cram Engine is a skill for students with one or two weeks before an exam. The user names the subject, the textbook, the question formats and the topics the lecturer marked. Then a four-stage pipeline takes over. Stage one turns the topic list into a knowledge tree where every node can be explained in about five minutes. Stage two teaches one node at a time: a familiar situation first, then the theory, at most three blocks, with a pause every three topics. Stage three runs test questions in real exam formats: multiple choice, cases, definitions, short and long answers. Stage four finds the root cause of each mistake and patches only the failures. It rests on six learning-science principles: load control, elaboration, the generation effect, retrieval practice, spacing and self-monitoring. It suits subjects with concept-based exam tickets best.
Who it is for. University students with an exam a week or two away and scattered notes who need a guide rather than a plain answer chat.
Good fit when
- Exam tickets consist of definitions, short answers and essays and there is no time to structure them.
- The lecturer flagged required topics and they must be worked through in order.
- After a mock test you want to see which mistake you make: mixing terms or failing to apply them.
- A course in medicine, law or networks needs a review without drowning in volume.
Not a fit when
- The exam is mostly computation, proofs and drawings: the skill will not drill them.
- You have no idea what will be tested: without lecturer hints the skill turns into an ordinary chat.
- You need a Russian-language skill: the repository texts and examples are in Chinese.
Example request
Use cram-engine to help me prepare quickly for an organizational behavior exam: tickets, textbook and topics attached.Limitations
The result depends on the quality of your input notes: if you do not know what the exam will cover, the skill behaves like a plain Q and A. It handles calculation tasks at the concept level and you still solve them yourself. The README, examples and prompts are written in Chinese; in another conversation language the agent teaches in that language, but the skill wording stays Chinese. The four-stage pipeline takes many messages, so context grows in one long dialog.
How to disable. Delete the `cram-engine` directory from your agent skills folder or use the matching command of your skills manager.
FAQ
How do I install Cram Engine?
Run `npx skills add https://github.com/liuliu667/cram-engine --skill cram-engine`. You can also send it to the agent in chat and let it install itself.
How do I start a session?
Tell the agent something like: use cram-engine to help me cram a subject. It will ask for the subject, textbook and exam scope.
Which subjects does it fit?
Subjects with definition, short and essay questions: economics, humanities, law, and the conceptual parts of medicine, networks and operating systems.
What does the testing stage do?
It asks questions in the lecturer formats: multiple choice, cases, situations and essays. Required topics appear in every format, extra ones are sampled.
What happens to mistakes?
The skill identifies the cause: mixed concepts, wrong logic, unmemorized terms or failure to apply. It explains differently and does not repeat the earlier text.
Discussion
No questions or reviews yet.
Related
SkillsAn exam prep skill built on your own lectures and problems: the agent goes through the course chapter by chapter with citations
A skill for exam and contest prep: topic checklist, day-by-day plan, daily lessons and a mistake notebook built from real answers.
A tutoring skill for STEM and AI: it finds the knowledge gap first, then teaches one step and checks understanding
A tutor for any subject: diagnosis, a study plan, project tasks and mastery checks, with the results saved to Markdown files
A skill that converts course materials into clean Markdown and prepares exam notes, practice questions and mistake analysis.
A learning engine for the agent: first-principles curricula, free-recall checks, FSRS-scheduled memory and explorable materials
README in short
The README explains which courses the skill suits and splits them into concept-based, mixed and computation-heavy groups. It then describes the four stages, the six learning principles and answers to common questions. The pipeline is built from stage markdown files in the `stages/` folder.
README badge
Original SKILL.md
--- name: cram-engine description: 期末速成引擎——一个懂学习科学的 AI 老师。基于认知负荷理论、精细加工、生成效应、检索练习，通过拆解→讲授→检题→补漏四阶段流水线，把课程重点变成交互式高效学习会话。一行命令启动，拯救大学生期末。 --- # 期末速成引擎 (Cram Engine) 期末速成学习工具。基于学习科学原理（认知负荷理论、精细加工、生成效应、检索练习），通过四阶段流水线将课程重点转化为个性化高效学习会话。 ## 触发条件 当用户提到以下任一情况时使用此 skill： - 期末速成、期末复习、考试冲刺、考前突击 - 使用 `/cram` 命令 - 明确表示要在短时间内备考某门大学课程 - 提到"帮我速成XX课" ## 命令 | 命令 | 行为 | |------|------| | `/cram <课程名> start` | 从阶段1开始完整学习流程 | | `/cram <课程名> resume` | 从进度文件中断处继续 | | `/cram <课程名> status` | 显示当前进度摘要（当前阶段、已学/未学知识点、错题数、顽固点数） | | `/cram <课程名> retry <知识点>` | 对指定知识点重新执行阶段2-3-4微型流程：重讲（四步教学）→ 重测（2道新题，至少一道题型不同于原题）→ 补漏（答错则诊断+换讲法+重测+顽固判定）。结果追加到进度文件，不覆盖原记录 | | `/cram <课程名> summary` | 输出考前清单：must_know 考点掌握状态（✅已掌握 / ⚠️顽固点 / ❌ 未纠正）、顽固点汇总、建议考前复习的知识点列表 | ## 文件路径约定 引擎文件（阶段指令、配置模板）位于本 SKILL.md 所在目录： - 阶段指令：`<skill_dir>/stages/` - 配置模板：`<skill_dir>/configs/` 用户数据（个人课程配置、学习进度）存储在用户主目录下： - 课程配置：`~/.cram-engine/configs/<课程名>.yaml` - 进度文件：`~/.cram-engine/progress/<课程名>-progress.md` - 首次使用时自动创建 `~/.cram-engine/` 目录结构 > `<skill_dir>` 指当前加载的本 `SKILL.md` 所在目录，阶段指令与配置模板均相对于该目录定位。