Install
$ agentstack add skill-cabbage2000-lab-textbook-writer-skills-textbook-exercises ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
textbook-exercises
为 STEM 教材生成三类题(示范例题/引导练习/独立习题),每道题带 Bloom 层级标注,每个计算答案都经过真算验证才输出。设计与验证规范的权威来源是 [references/exercise-design.md](references/exercise-design.md)。
何时不触发
- 文科论述题、案例分析题(v1 不支持;被要求时说明属 v2 范围,可改出概念辨析题替代)
- 组卷/试卷排版(本 skill 只产题目内容)
- 非 STEM 内容的一般性问答
两种调用模式
- 被 textbook-chapter 调用(常规):输入 = 契约
{章上下文, 题目计划, 术语表}(字段定义见 [../textbook/references/handoff-contract.md](../textbook/references/handoff-contract.md) 第 2 节)。按题目计划逐条生成,不多不少。 - 独立触发:一轮问清
{主题, 读者水平, 三类题各要几道, 期望 Bloom 层级分布}(用户已说明的不重复问),术语自拟并保持内部一致;编号中的章号默认取 1,用户指定章号则从之。
生成流程(每道题走完 1–4 才算完成一道)
- 设计题面:按 [references/exercise-design.md](references/exercise-design.md) 第 4 节对应题类的要点设计,对照题目计划的主题与 Bloom 层级;层级判定标准查 [../textbook-outline/references/bloom-levels.md](../textbook-outline/references/bloom-levels.md) 第 4 节(同题多问取最高层,歧义就低不就高)。
- 同步写完整解答:示范例题 = 逐步推导每步一行理由;引导练习 = 提示阶梯(提示 1 方向 → 提示 2 关键步骤)+ 完整解答;独立习题 = 参考答案或含关键中间结果的解题路径。禁止先攒题后补答案。
- 真算验证(红线,不可省略):按 exercise-design.md 第 2 节选验证方式——
- 数值计算/符号推导类:写一段与解答路径不同的 Python(sympy/numpy)复算代码,用 Bash 实际执行,比对结果;
- 代码题:实际运行代码核对输出;
- 证明题:逐步核验每步依据(引用的定义/定理是否成立且已在教材前文出现);
- 概念辨析:核对定义来源;
- 当前环境无法执行 Python 时:手算两遍交叉核对,验证状态注明"手算交叉核对"。
- 处理验证结果:复算一致 → 标
✅ 已验证();不一致 → 修题或修解答后重新验证;连续 2 次验证失败 → 弃题换题。确实无法机器验证且手算不可行(如开放性数值实验)→ 标⚠️ 需作者确认(),绝不假装已验证。
输出
- 按契约逐题给出
{编号, 类型, 题干, 解答或提示, Bloom层级, 验证状态}; - 正文格式遵循 exercise-design.md 第 6 节(编号规则
例 N-M/习题 N-M,引导练习计入"例"序列加后缀); - 验证用的复算代码随题附上(放在"验证"行后的代码块中),供作者复查;
- 术语用词严格对照传入的术语表;独立触发时保持自拟术语的内部一致。
纪律(违反任何一条即返工)
- 绝不输出未验证的计算答案——"应该没错"不是验证;
- 开放题给参考要点并标注"开放题",不编造标准答案;
- 每题必有 Bloom 标注,漏标即不完整;
- "留给读者作为练习"式悬空 = 违规(exercise-design.md 第 3 节第 6 条);
- 出题偏离题目计划的主题或层级时,须在输出中注明偏离原因(如"该主题在应用层无自然题型,改为分析层"),由调用方决定是否接受。
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: cabbage2000-lab
- Source: cabbage2000-lab/textbook-writer-skills
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.