Learning Cards (Spaced Repetition)

Spaced-repetition flashcard system backed by Feishu Bitable (multi-dimensional table). Create flashcards from any book or knowledge base, quiz the user interactively, track progress with 16 fields, and auto-schedule reviews using the 85% Rule. 基于飞书多维表格的间隔复习学习卡片系统。从任何书籍/知识体系生成学习卡片， 通过对话式问答练习，16 个字段追踪学习进度，85% 规则动态调整难度。 Use when / 使用场景: (1) User wants to study a book/course with flashcards / 用户想用卡片学习一本书或课程 (2) User says "继续学习", "来几张卡片", "continue studying", "quiz me" (3) User asks to create learning cards from a book or document / 用户要求从书籍或文档生成学习卡片 (4) User asks about learning progress, weak points, or review schedule / 查看学习进度、薄弱点、复习计划 (5) User mentions "学习卡片", "间隔复习", "spaced repetition", "flashcards"

Learning Cards

Spaced-repetition flashcard system on Feishu Bitable.

基于飞书多维表格的间隔复习学习卡片系统。

Full design details: references/system-design.md

完整设计文档：references/system-design.md

Prerequisites

This skill requires Feishu (Lark) Bitable access. Ensure the following before use:

本技能依赖飞书多维表格。使用前请确认以下条件：

1. OpenClaw Feishu plugin configured — The openclaw-lark plugin must be installed and connected to a Feishu tenant. See OpenClaw Feishu setup docs.

   OpenClaw 飞书插件已配置 — 需安装 openclaw-lark 插件并连接到飞书租户。

2. User OAuth authorization — The user must grant Bitable read/write permissions when prompted (standard Feishu OAuth flow, handled automatically by OpenClaw).

   用户 OAuth 授权 — 用户需在首次使用时授权多维表格读写权限（OpenClaw 自动处理）。

3. Bitable app scope — The Feishu app needs scopes: bitable:app, bitable:app:readonly. These are standard in the OpenClaw Feishu plugin.

   多维表格权限范围 — 飞书应用需开通 bitable:app、bitable:app:readonly 权限（OpenClaw 飞书插件标配）。

No additional API keys or credentials are needed — authentication is handled by the OpenClaw Feishu plugin's existing OAuth infrastructure.

无需额外 API 密钥 — 认证由 OpenClaw 飞书插件的 OAuth 基础设施统一处理。

Setup

Create a Bitable app with one table containing these fields.

创建一个飞书多维表格，包含以下字段。

Content fields / 内容字段

Field	Type	Note
概念名	Text (primary)	Card title / 卡片标题
阶段	SingleSelect	Major phase / 大阶段
站点	SingleSelect	Chapter/unit / 章节
类型	SingleSelect	e.g. 核心概念 / 隐喻 / 转变工具 / 关键问题
正面（问题）	Text	Question / 问题
背面（答案）	Text	Answer / 答案
我的理解	Text	Learner's own notes / 学习者笔记

Tracking fields / 追踪字段

Field	Type	Note
掌握程度	SingleSelect	未学习 / 初步了解 / 基本掌握 / 深度理解
学习顺序	Number	Sequential order for first pass / 首轮学习顺序
首次学习时间	DateTime	First study timestamp (ms) / 首次学习时间戳
上次复习	DateTime	Last study timestamp (ms) / 上次学习时间戳
下次复习时间	DateTime	Next review date (ms) / 下次复习日期
学习次数	Number	Total attempts / 总答题次数
连续答对次数	Number	Streak — drives interval / 连对次数，决定间隔
答错次数	Number	Cumulative errors / 累计答错
最近得分	Number	Last score (1-5) / 最近得分

Three Learning Phases

Phase 1: Sequential coverage / 顺序覆盖

First pass through all cards in order by 学习顺序. Goal: build a mental map of the entire knowledge base.

按学习顺序首轮过一遍，建立全局地图感。

Phase 2: Spaced review / 间隔复习

After first pass. Schedule by 下次复习时间. Mix in remaining new cards.

首轮完成后按复习时间调度，穿插新卡。

Phase 3: Mastery & transfer / 融会贯通

Cross-topic quizzes, scenario-based questions, reverse prompts (give answer → guess concept).

跨章节混合抽问、情境题、反向提问（给答案猜概念）。

Quiz Flow

text

1. Read all cards from Bitable
2. Compute recent accuracy (last 5 scores) → determine new:review ratio
3. Select cards:
   - Review: 下次复习时间 <= today, sorted by overdue days
   - New: 学习次数 = 0, sorted by 学习顺序
4. Mix per ratio, present one at a time
5. After each answer: score → show answer → update Bitable

Card presentation format / 出题格式

text

第 X 题（复习/新卡）
> [概念名] | [阶段] | [站点] | [类型]
[正面问题]

User can say "讲讲" to skip answering and see the explanation directly (score = 0).

用户可说"讲讲"跳过作答，直接看答案（计 0 分）。

85% Rule — Dynamic Ratio

Target ~85% accuracy for optimal learning speed (Wilson et al., Nature Communications, 2019).

目标保持 85% 正确率以获得最优学习速度。

Recent accuracy	New:Review	Logic
> 90%	3:1	Too easy, add new / 太轻松，加新卡
80–90%	2:1	Optimal zone / 最优区间
70–80%	1:1	More review / 多复习
< 70%	1:2	Consolidate first / 先巩固

Accuracy = proportion of scores >= 4 in last 5 attempts.

正确率 = 最近 5 次中得分 >= 4 的比例。

Spaced Review Algorithm

Interval based on consecutive correct answers (连续答对次数):

Streak	Interval
0	1 day
1	1 day
2	3 days
3	7 days
4	14 days
5	30 days
6+	90 days

Post-Answer Update

After each answer, immediately update the card record:

每题答完立即更新卡片记录：

text

学习次数 += 1
最近得分 = score
上次复习 = today
首次学习时间 = today  (if null)

if score >= 4:       连续答对次数 += 1
elif score <= 2:     连续答对次数 = 0; 答错次数 += 1 (if score > 0)
else (score == 3):   streak unchanged

intervals = [1, 1, 3, 7, 14, 30, 90]
下次复习时间 = today + intervals[min(streak, 6)] days  (if score >= 4)
               today + 1 day                           (otherwise)

掌握程度:
  streak >= 5 → 深度理解
  streak >= 3 → 基本掌握
  studied    → 初步了解

Card Generation Guidelines

When creating cards from a book:

从书籍生成卡片时：

• 2–4 cards per chapter for conceptual books / 概念类书籍每章 2-4 张
• 1–2 per tool/API for technical books / 技术类每个工具 1-2 张
• Target 30–50 cards total / 总量 30-50 张
• Questions should be specific, not vague / 问题要具体
• Answers: 150–300 chars, use ①②③ for key points / 答案 150-300 字，序号标注

User Commands

Command	Action
继续学习 / continue	Default 3 cards, dynamic ratio / 默认 3 张，动态配比
来 N 张 / give me N	Specific count / 指定数量
讲讲 / explain	Show answer directly / 直接看答案
学习进度 / progress	Show overall stats / 显示整体进度
薄弱点 / weak points	List cards with most errors / 列出高错卡片