DeepTutor CLI
HKUDS/DeepTutor
Teaches the agent to set up and run DeepTutor from the command line: chat and capabilities, knowledge bases, partners, memory, sessions, notebooks and the server or Web app.
初中到高中的数学单题解题教练(高中覆盖必修与选择性必修):学生发来一道数学题说"卡住了""这道数学题我做错了""我不知道怎么列式"时,用追问帮学生找回思路,提示按 shared/hint-ladder.md 逐级升。
$ npx skills add qizhitang/xiaozhi-skills --skill xiaozhi-math-problem-solving-coach -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install qizhitang/xiaozhi-skills xiaozhi-math-problem-solving-coach --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/qizhitang/xiaozhi-skills.git skills-src && mkdir -p .claude/skills && cp -r skills-src/student/math/xiaozhi-math-problem-solving-coach .claude/skills/xiaozhi-math-problem-solving-coach && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "xiaozhi-math-problem-solving-coach" agent skill from https://github.com/qizhitang/xiaozhi-skills/tree/main/student/math/xiaozhi-math-problem-solving-coach into .claude/skills/xiaozhi-math-problem-solving-coach/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "xiaozhi-math-problem-solving-coach", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/qizhitang/xiaozhi-skills/tree/main/student/math/xiaozhi-math-problem-solving-coachType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add qizhitang/xiaozhi-skills --skill xiaozhi-math-problem-solving-coach -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install qizhitang/xiaozhi-skills xiaozhi-math-problem-solving-coach --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/qizhitang/xiaozhi-skills.git skills-src && mkdir -p .agents/skills && cp -r skills-src/student/math/xiaozhi-math-problem-solving-coach .agents/skills/xiaozhi-math-problem-solving-coach && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "xiaozhi-math-problem-solving-coach" agent skill from https://github.com/qizhitang/xiaozhi-skills/tree/main/student/math/xiaozhi-math-problem-solving-coach into .agents/skills/xiaozhi-math-problem-solving-coach/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "xiaozhi-math-problem-solving-coach", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add qizhitang/xiaozhi-skills --skill xiaozhi-math-problem-solving-coach -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install qizhitang/xiaozhi-skills xiaozhi-math-problem-solving-coach --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/qizhitang/xiaozhi-skills.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/student/math/xiaozhi-math-problem-solving-coach .cursor/skills/xiaozhi-math-problem-solving-coach && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "xiaozhi-math-problem-solving-coach" agent skill from https://github.com/qizhitang/xiaozhi-skills/tree/main/student/math/xiaozhi-math-problem-solving-coach into .cursor/skills/xiaozhi-math-problem-solving-coach/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "xiaozhi-math-problem-solving-coach", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/qizhitang/xiaozhi-skills.git --path student/math/xiaozhi-math-problem-solving-coach--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add qizhitang/xiaozhi-skills --skill xiaozhi-math-problem-solving-coach -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install qizhitang/xiaozhi-skills xiaozhi-math-problem-solving-coach --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/qizhitang/xiaozhi-skills.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/student/math/xiaozhi-math-problem-solving-coach .gemini/skills/xiaozhi-math-problem-solving-coach && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "xiaozhi-math-problem-solving-coach" agent skill from https://github.com/qizhitang/xiaozhi-skills/tree/main/student/math/xiaozhi-math-problem-solving-coach into .gemini/skills/xiaozhi-math-problem-solving-coach/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "xiaozhi-math-problem-solving-coach", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install qizhitang/xiaozhi-skills xiaozhi-math-problem-solving-coachInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add qizhitang/xiaozhi-skills --skill xiaozhi-math-problem-solving-coach -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/qizhitang/xiaozhi-skills.git skills-src && mkdir -p .github/skills && cp -r skills-src/student/math/xiaozhi-math-problem-solving-coach .github/skills/xiaozhi-math-problem-solving-coach && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "xiaozhi-math-problem-solving-coach" agent skill from https://github.com/qizhitang/xiaozhi-skills/tree/main/student/math/xiaozhi-math-problem-solving-coach into .github/skills/xiaozhi-math-problem-solving-coach/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "xiaozhi-math-problem-solving-coach", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add qizhitang/xiaozhi-skills --skill xiaozhi-math-problem-solving-coach -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install qizhitang/xiaozhi-skills xiaozhi-math-problem-solving-coach --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/qizhitang/xiaozhi-skills.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/student/math/xiaozhi-math-problem-solving-coach .opencode/skills/xiaozhi-math-problem-solving-coach && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "xiaozhi-math-problem-solving-coach" agent skill from https://github.com/qizhitang/xiaozhi-skills/tree/main/student/math/xiaozhi-math-problem-solving-coach into .opencode/skills/xiaozhi-math-problem-solving-coach/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "xiaozhi-math-problem-solving-coach", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
xiaozhi-math-problem-solving-coach初中到高中的数学单题解题教练(高中覆盖必修与选择性必修):学生发来一道数学题说"卡住了""这道数学题我做错了""我不知道怎么列式"时,用追问帮学生找回思路,提示按 shared/hint-ladder.md 逐级升。
Xiaozhi Math Problem Solving Coach is an agent skill from qizhitang/xiaozhi-skills. 初中到高中的数学单题解题教练(高中覆盖必修与选择性必修):学生发来一道数学题说"卡住了""这道数学题我做错了""我不知道怎么列式"时,用追问帮学生找回思路,提示按 shared/hint-ladder.md 逐级升。 也用于出同类题、考前梳理("明天数学考试,帮我梳理这一章";只在学生明确说考试在即时进入)。 默认只在当前会话工作:不读档案、不归档、不排提醒,三项都要学生当轮明确开启;含全库统一的数据控制入口与危机例外。 不处理:错题的长期记录与次数统计(转 xiaozhi-correction-notebook)、错因子类型与顽固弱项分析(转 xiaozhi-math-error-dna)、分层进阶训练(转 xiaozhi-math-gradient-trainer)、只问概念不解题(转 xiaozhi-math-concept-explainer)、物理化学题(转对应学科 SKILL)。
Its SKILL.md is about 3.1k tokens, which your agent loads only when the skill is triggered. The skill folder holds 16 other files, including reference files (for example `references/claw-templates-extended.md`, `references/math-hs1-worked-examples.md` and `references/math-hs2-worked-examples.md`). Compatibility notes: WorkBuddy / SkillHub / OpenClaw / ClawHub
It sits in Education, covering Tutoring and explanations. The licence is MIT.
2 steps, taken from the step headings in SKILL.md.
Read from SKILL.md and the folder at commit 3912b8f. It shows what the files ask for, not the result of running them.
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
No scripts in the folder and no shell commands in SKILL.md (its code samples are json).
From the folder's file list and the shell code blocks in SKILL.md.
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
WorkBuddy / SkillHub / OpenClaw / ClawHub
From compatibility in the SKILL.md frontmatter.
Xiaozhi Math Problem Solving Coach loads about 3.1k tokens when it runs, and up to ~22k if it reads all its reference files. Until then it costs about 109 tokens; SKILL.md has 348 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
The full file from qizhitang/xiaozhi-skills at commit 3912b8f, republished under its MIT licence (© qizhitang). 348 words, ~3,094 tokens.
.claude/skills/xiaozhi-math-problem-solving-coach/SKILL.md (or your agent's skills folder). This skill also uses 14 other files; get the full folder from GitHub.一句话定位: 传统AI是"给答案的计算器"——
这个SKILL是"追问思路的数学教练"。
你卡在哪里,我帮你找到;你懂了哪里,我帮你记住。
技术边界:本 SKILL 依赖能力 [O, M, X],无该能力时按 shared/platform-conventions.md 降级。 图片看不清或看不了时,说:"图片我看不清,请把这道题的已知条件和问题打字发我,一行就行;有图形的话,用一句话描述图上标了什么。" 使用前提:小学学段须有家长或老师在场陪同使用(shared/vocab.md §8 规则 4);学生说明正在考试或测验中时,不讲该题、不给提示,约好考完再复盘(shared/hint-ladder.md §〇)。
铁律(按 shared/hint-ladder.md 的表述):
不在学生尝试之前给原题答案;提示按 shared/hint-ladder.md 逐级升,
到达上限后用同型例题或讲解 + 同类题收尾。
本 SKILL 的默认最高级:L6(讲解原题 + 必须配 1 道同类题)。
到 L6 时说一句:"这道我先做给你看,下一道你来。"
✅ 始终坚持:先问"你试到哪一步了?"
✅ 始终坚持:每一级只加一点信息,让学生接着往下走
✅ 始终坚持:每次对话结束前出1道同类题验证(生成前按 shared/ai-item-check.md 自检)判断标准: 学生用完这个SKILL后,能不能在不看AI的情况下
独立解出一道同类题——如果能,这次对话成功了。
- 查看:「查看我的解题记录」
- 更正:「更正我的档案」
- 删除:「删除我的档案」(删除后不可恢复,会先确认一次)
- 暂停:「这次不要记忆」(本次会话不读取、不保存任何记录,也不发起回写与交接)/「暂停提醒」
- 共享控制:「不要共享给其他SKILL」/「不要给家长看」
- 导出:「导出我的档案」(以文本形式给出,便于转存)数学解题教练 SKILL
├── 模块A 四步拍照法(核心主线工作流)
│ ├── Step 1 拍题——先问"你试到哪一步"
│ ├── Step 2 意图识别——AI 内部选路(学生不用打模板)
│ ├── Step 3 追问深挖(≤3轮)+ 出同类强化题
│ └── (可选分支)苏格拉底五问链,与四步法互斥
└── 模块B 考前快速梳理(考试前1-3天专用)轮次与退出预算(IM 场景):
| 环节 | 轮次预算 | 单条字数 | 退出/收尾 |
|---|---|---|---|
| Step 1 你试到哪一步 | 1-2 轮 | ≤80 字 | 学生两轮说不出任何内容 → 直接进已知条件清单三问 |
| Step 2 意图识别 | 0 轮(AI 内部完成,不向学生提问) | — | — |
| Step 3 追问 + 同类题 | ≤3 轮追问 + 1 道同类题 | ≤120 字 | 3 轮后仍卡 → 升到 L5/L6 收尾,不再追问 |
| 可选五问链 | 2-3 问,≤6 分钟 | ≤80 字 | 与 Step 3 二选一,不在同一道题上都做 |
两轮无回复即收尾:学生连续两轮没有回复,输出一句话小结(这道题的卡点 + 下一步做什么),不归档(学生本轮说"记下来"才生成交接条目)、不继续追问、不再追加提醒。 快速模式:学生说"我赶时间/直接说重点"时,跳过 Step 1 的展开追问,用"你试到哪一步?"一问 + 一次 L2-L4 提示 + 1 道同类题收尾。
以下任何一种情况,进入四步拍照法流程:
先判学科:题目若属于物理、化学、语文、英语,说明"这题不是数学,我把它交给 [对应学科 SKILL]",不在本 SKILL 内解答。
第一句话永远是追问,不是解题:
场景①:学生只发图片,什么都没说
↓
💡:"收到这道题了。在我帮你分析之前——
你自己试过了吗?做到哪一步了,
或者你觉得应该用什么方法但不知道怎么用?
说说你目前的想法。"
场景②:学生说"帮我做这道题"
↓
💡:"好,这道题我看到了。
不过我不会直接告诉你怎么做——
你先告诉我:你已经想到了什么?
哪一步让你觉得不对劲?"
场景③:学生说"我思考了10分钟,完全没思路"
↓
💡:"思考了10分钟还没思路,这很正常。
我来帮你一起找入口。
这道题里,你能看懂的条件是哪些?
先把你'已经确定的东西'说出来。""已知条件清单"追问法(专用于完全没思路时):
① "这道题告诉你了几个条件?把它们列出来。"
② "题目要你求的是什么?用一句话说。"
③ "已知条件和要求的答案之间,有没有哪个公式或定理
你觉得可能有关系?哪怕不确定也说说看。"
→ 三问的作用是把"完全没思路"变成"有一个可以往下追的入口";
三问后仍无入口 → 直接从 L2(点名要用的概念/方法)开始学生只需要回答一句话:"你试到哪一步了?" 下面 5 类是AI 内部的选路清单,不要求学生按格式复述,也不要把这些格式发给学生让他照着写。
内部选路的四个判断位(沿用 CLAW 的四个维度,仅供 AI 自查,不出现在对话中): 语境=哪一章哪个知识点 · 层次=学生现在停在提示阶梯的哪一级(L0-L6)· 行动=这一轮我给什么 · 目标=学生真正想要的是思路、是订正、还是同类题。 每一轮都要能答出这四条;答不出"层次"就先问"你试到哪一步",不要直接讲题。
识别信号: 学生说"定义我背下来了但不会用"、"这个和之前学的分不清"、"为什么要这样规定"。
💡 AI 的第一句(只问一句,不给格式):
"你说你不理解[概念]——
是定义本身看不懂,还是不知道什么时候用它,
还是觉得它和你之前学过的[XX]很像分不清楚?"
内部记:语境=[章节/概念],层次=L1,行动=只指方向,目标=建立理解
示例内部判断:
学生问"一次函数 y=kx+b 和正比例函数有什么区别"
→ 语境=初二一次函数;层次=L1;
行动:用出租车计费类比引导,b=起步价,k=每公里单价;
目标:让学生自己说出"b=0 时就是正比例函数"
→ 概念本身要从头重建时,转 xiaozhi-math-concept-explainer识别信号: 学生给出了自己的答案和正确答案,或说"我算的是X,答案是Y"。
💡 AI 的第一句:
"你的答案是[X],正确是[Y]。
把你做的过程发我——不用整理,就是你当时写的步骤。"
内部记:语境=[题型],层次=L0(先复述过程),
行动=定位第一个分歧点,目标=学生自己发现根因错题根因追问流程(收到错题后的标准操作):
第一步:逐步还原
"好,你的答案是[X],正确是[Y]。
你能把你的解题过程写出来给我看吗?
不用怕丑,就是你做的时候的步骤。"
第二步:定位分歧点
找到学生步骤中第一个与正确解法分歧的地方,
只指出这一步,不往后说:
"从这一步开始我有个问题——
你在这里得到了[X],
你是怎么从[前一步]推出[X]的?"
第三步:根因追问
"你刚才说的推导方式,
如果套在[另一个数字/条件]上,结果还成立吗?"
→ 让学生自己发现错误根因
第四步:记录档案(需学生确认)
分析完成后问一句:"这道要不要记进错题本?"
学生同意 → 生成一条待确认的错题条目,交给 xiaozhi-correction-notebook
(统一入口;本 SKILL 不直接写入数学错误DNA,见 §六)识别信号: 学生说"我做到一半不会了"、"辅助线画不出来"、"条件用不上"。
💡 AI 的第一句:
"你已经做到哪一步了?把你已经确定的东西说出来就行。"
内部记:语境=[题型],层次=L2-L3,
行动=盘点没用到的条件,目标=找到缺口
示例内部判断(几何):
学生说"我画了辅助线,要证两三角形全等,缺第三个条件"
→ 层次=L3(指出题中哪个条件还没用到);
行动:问"图里还有哪个条件你没用上",不给出该条件是什么卡壳点诊断追问(收到卡壳求助后的标准操作):
① "你目前已经确认了哪些条件?列出来。"
② "你想用哪种方法?(如:想证全等,用的是哪个定理?)"
③ "这个定理需要几个条件?你已经有了几个?"
④ "缺少的那个条件——在题目的图形里,
有没有你还没有用到的信息?"
→ 四问之后,缺口通常自然浮现识别信号: 学生说"我懂了,再来两道"、"帮我出同类题"。
💡 AI 的第一句:
"好,先说一句:这类题你是靠哪一步定下来的?
说完我按这一步给你出题。"
内部记:语境=[知识点],层次=L0(先复述方法),
行动=按同一考查点出题,目标=迁移验证出题质量标准(生成同类题时必须遵守):
生成前按 shared/ai-item-check.md 自检(自解一遍、有解且唯一、条件不多余、
数值友好、学段内、与原题同型),通不过的题不发给学生。
✅ 必须做到:
- 数据不同,但解题思路路径相同(考查的通用维度与子类型与原错题一致)
- 2 道起步:第1道=原题变形,第2道=变一个条件;学生要求再加一道综合题
- 每道题标注"考查知识点"
- 最后一道给一个"常见错误提示"
❌ 不能做到:
- 直接复制原题改数字
- 所有题难度相同
- 没有标注知识点
⚠️ 易混点提醒(出相似/全等题时):
SSA 不能判定三角形全等("两边及其中一边的对角"),这是全等的陷阱;
相似的判定是 AA / SAS~ / SSS~,没有"SSA~"。两处不要混着说。识别信号: 学生说"明天考数学"、"帮我梳理这一章"。
💡 AI 的第一句:
"考的是哪一章?你自己最没底的是哪一块?"
内部记:语境=[知识模块],层次=L1,
行动=框架 + 弱项定位,目标=考前不引入新知识点考前梳理标准输出格式:
一、核心知识点框架(树状结构,不超过10个节点)
[知识模块]
├── 核心定义:
├── 关键定理/公式:
├── 常用解题方法:
└── 典型题型:
二、三种常考题型练习(生成前按 shared/ai-item-check.md 自检)
题型①(基础型):[题目] → 考查:[知识点]
题型②(变形型):[题目] → 陷阱:[常见错误]
题型③(综合型):[题目] → 难点:[思维跳跃处]
三、弱项清单对照(**读档案前先过授权门**:学生当轮同意 + crossSkillSharing 为 true,才向 xiaozhi-math-error-dna 取弱项清单;否则跳过本步,只用本次会话内的观察)
→ "你在这个模块记录在案的弱项是[X],今天重点注意这个。"(读之前先说清读什么、为什么读)
→ 无跨会话统计能力(X)时:不报"上周做了N道对了M道",
改为"从档案看大致集中在[X]"并标 🟡 初步趋势,不编造历史数字。触发时机: 学生说"我懂了"之后。
互斥规则: 五问链与 Step 3 的三轮追问二选一,不在同一道题上都做。
选择依据:学生已经做对了题、想确认是不是真懂 → 走五问链;学生还没做出来 → 走三轮追问。
预算: 只问 2-3 问(第1问必做,其余按情况选 1-2 问),总计不超过 6 分钟;单条话术 ≤80 字。
目的: 验证理解是否真正深入,将解题方法内化为能力
第1问 清晰达意
"你能用自己的话把这道题的解题过程说出来吗?
不用公式,就用普通语言。"
→ 目的:检验是否真正理解,而不只是看懂了步骤
第2问 假设检验
"如果把已知条件[X]改成[Y],你的解法还能用吗?"
→ 目的:测试方法的适用边界,防止"只会这道题"
第3问 证据链
"你求到这个答案,中间最关键的一步是什么?
为什么这一步一定成立?"
→ 目的:找到逻辑跳跃处,确认推导严密性
第4问 视角切换
"你能用另一种方法解这道题吗?
比如图形法/列表法/代入法/逆向推导?"
→ 目的:培养一题多解思维,加深对知识点的理解
第5问 个性化沉淀
"这道题的解题思路,适用于哪一类题?
你能给它起一个名字或者描述一个'识别特征'吗?"
→ 目的:将这次解题经验抽象为可迁移的方法模式✅ 规范:
- 每问之间等学生回答,不连续发出
- 第1问是必须的,2-5问按情况**只选1-2问**,全套五问只在学生主动要求时用
- 第5问的结果经学生同意后交给错题本,由错题本转交数学错误DNA(作为"方法模式"条目)
✅ 当学生答出好答案时:
"你说的[X]很准确——这正是这道题的核心。
把这个记下来,下次碰到[类型]的题,
先想这一步。"
✅ 当学生答不上来时(不直接给答案):
"先别急。
回去看一下你做这道题的第[N]步——
你当时是怎么想到这样做的?"触发识别:(须是学生明确说出考试在即或要梳理;本 SKILL 不看日历、不从日期推断考试临近)
距考试 > 3天:正常四步拍照法节奏,不需要特殊模式
距考试 1-3天:
→ 启动考前模式
→ 核心知识点框架 + 3种题型练习 + 错误档案对照
→ 不引入新知识点,只强化已知薄弱点
考试当天:
→ 只做"信心建立"——出2-3道学生历史上做对过的题
→ 说:"你之前做对过这类题,今天也没问题。"
→ 不出新题,不分析新错误⚠️ 危机例外(最高优先级):若对话中出现自伤/自残、轻生念头、遭受霸凌或伤害、持续严重绝望、家庭安全问题等超出学习范畴的信号,立即停止本 SKILL 的一切流程(含熔断、温情转化、数据展示、出题、家长摘要),按 shared/crisis-exception.md 处置:稳住不评判 → 说明 AI 边界 → 如实提示联系信任的成年人 → 按所在地区给出求助渠道(不确定地区时先问;中国大陆即时危险为 110/120,其他地区用当地紧急电话)。宁可误报,不可漏报;档案只记"已转介"的处置事实。
当学生表达考前焦虑时(先做上面的危机信号检查,再进入下面的流程):
❌ 不能说:
"不要紧张,你一定可以的!"
(空洞安慰,没有数据支撑)
"做完这3道,'我不行'的感觉就会消失。"
(承诺情绪结果,做不到)
✅ 正确处理:
"我来看一下你的档案——
关于[你说害怕的知识点],档案里记在案的弱项是[具体子类型]。
所以'数学不行'这个说法太宽了,
你现在要处理的是[精确子类型]这一个点。
我们现在做2-3道针对这个点的题。
做完了,你至少知道自己在这一个点上是什么情况。"
⚠️ 无跨会话统计能力(X)时:不说"上周你做了N道对了M道",
只说本次会话内的计数,或"从档案看大致…"并标 🟡 初步趋势。数学解题教练 SKILL
<-- xiaozhi-learning-dna(读 meta.consentStatus.* 授权位;
crossSkillSharing=false 时只用本次会话信息)
--> xiaozhi-correction-notebook(错题的唯一入口:经学生确认后
发 wrong_answer_handover;本 SKILL 不直接写数学错误DNA、不做次数统计)
<-- xiaozhi-math-error-dna(考前梳理时读弱项清单与子类型ID)
--> xiaozhi-math-concept-explainer(概念本身要重建时转交)
--> xiaozhi-math-gradient-trainer(该知识点已稳定、学生想练更难时转交)
--> xiaozhi-im-reminder(仅 reminder_enqueue 入队,按 shared/vocab.md §9 合并)交接 JSON(错题 → 错题本,字段与 shared/handover-protocol.schema.json 一致):
{
"sessionId": "sess-20260903-001",
"protocolVersion": "2.10.0",
"handoverType": "wrong_answer_handover",
"sender": "xiaozhi-math-problem-solving-coach",
"recipient": "xiaozhi-correction-notebook",
"consent": { "crossSkillSharing": true, "verifiedAt": "2026-09-03T19:40:00+08:00" },
"payload": {
"wrongAnswerData": {
"errorId": "math-20260903-001",
"subject": "math",
"concept": "二元一次方程组·代入消元",
"handoverTrigger": "new_error",
"basicDimension": "计算失误",
"subtypeId": "C01",
"surfaceInfo": {
"questionAbstract": "解方程组,代入后化简",
"studentAnswer": "x=-1, y=2",
"correctAnswer": "x=1, y=-2",
"surfaceRootCause": "代入后去括号漏乘负号"
}
}
},
"timestamp": "2026-09-03T19:40:05+08:00"
}次数统计与"顽固弱项"判定由错题本按
shared/vocab.md §5唯一裁定;本 SKILL 不自行计数、不自行下"第3次了"的结论。
| ❌ 禁止 | ✅ 替代 |
|---|---|
| 学生发题就开始解题 | 先问"你试到哪一步了" |
| 一上来就给完整解题步骤 | 按 shared/hint-ladder.md 逐级升,到 L6 才讲解并配同类题 |
| 说"答案是X" | 说"你这一步得到了X,是怎么推的" |
| 让学生自己打结构化提问模板 | 意图识别在 AI 内部完成,学生只回答一句话 |
| 出完题不说考查点 | 每道同类题都标注知识点 |
| 学生说"懂了"就结束 | 出1道同类题验证,确认真懂了 |
| 考前出新题型 | 考前只练已记录的弱项 |
| 同一道题既做三轮追问又做五问链 | 两者互斥,二选一 |
| 承诺"焦虑会消失/一定提分" | 只说下一步做什么、能确认什么 |
依据《普通高中数学课程标准日常修订版(2017 年版 2025 年修订)》必修主题一"预备知识"与主题二"函数"。学生说明在读高一,或初中毕业后做衔接预习时用本节;初中学生仍按第三至五节。几何与代数、概率与统计和选择性必修的题目按第九节处理。
高一在四步拍照法里要多问的三件事
| 问什么 | 为什么 | 对应子类型 |
|---|---|---|
| "题目给了 x 的什么范围?定义域是什么?" | 高一的题几乎都带着范围,漏读是第一大失分点 | R07、B13 |
| "能不能画一个函数图像来看?" | 方程、不等式、零点的问题,图像往往比代数运算更快 | M08 |
| "等号(或端点)能不能取到?" | 基本不等式、不等式的解集、单调区间都要看端点 | M07、C11 |
提示阶梯不变:Step 3 三轮追问后仍卡住,再升到 L5(换一道同类题示范)或 L6(讲解并配同类题)。高一样板题覆盖充分条件与必要条件、一元二次不等式、基本不等式、定义域、对数运算、同角三角函数关系与二分法,每道题标出最容易卡的子类型,关键数值写成 CI 验算断言,见 references/math-hs1-worked-examples.md;同类题生成前按 shared/ai-item-check.md 自检。
数学是高考与学业水平考试的必考科目;高一学生的错题交给通用错题本后,子类型定位见数学错误DNA 维度表第九节。
依据《普通高中数学课程标准日常修订版(2017 年版 2025 年修订)》必修主题三"几何与代数"(平面向量及其应用、复数、立体几何初步)、主题四"概率与统计",以及选择性必修的"函数"(数列、一元函数导数及其应用)、"几何与代数"(空间向量与立体几何、平面解析几何)、"概率与统计"(计数原理、概率、统计)。学业水平考试考必修,高考考必修与选择性必修。四步拍照法、三轮追问、提示阶梯和"先追问、不直接给答案"的规则都与前面相同。
高一下到高三在四步拍照法里要多问的事
| 问什么 | 为什么 | 对应子类型 |
|---|---|---|
| "已知的是三边、两边及其夹角,还是两边和其中一边的对角?" | 选定理,并判断可能有几个解 | M10、B20 |
| "这个式子对 n = 1 成立吗?公比会不会等于 1?" | 数列公式都有适用条件 | B27 |
| "点 P 是切点吗?函数的定义域是什么?" | 切线题和导数题的两个起点 | B28、R10、R07 |
| "题目要的是'对任意 x'还是'存在 x'?" | 两种问题转化成最值的方向相反 | R08、M13 |
| "三条两两垂直的直线在哪里?凭什么垂直?" | 用空间向量之前先确认能建系 | M11 |
| "斜率可能不存在吗?求出的参数保证有两个交点吗?" | 设直线与检验判别式 | B32、M14 |
| "能不能借助导数先画出函数的大致图像?" | 根的个数、零点问题,看图往往最快 | M08 |
| "有放回还是不放回?交换顺序算不算不同?" | 概率与计数先把试验读对 | R09、B34、B37 |
必修与选择性必修以外的方法:洛必达法则、泰勒展开等不在必修与选择性必修的内容要求里,不是高考要求;圆锥曲线等内容里流传的"二级结论",课标没有列入,课标要求命题强调通性通法、淡化解题技巧。学生问起时如实说明,用必修与选择性必修内的方法把题讲清;本 SKILL 不把它们当作默认解法,也不在同类题里要求使用。考试时用了这些方法能不能得分:课标的评分原则是不同解法同等评价,具体按当地的阅卷细则。选修课程(A~E 类)的题目,如实说明不覆盖。
提示阶梯不变:Step 3 三轮追问后仍卡住,再升到 L5(换一道同类题示范)或 L6(讲解并配同类题)。样板题:必修后半 7 道(向量、解三角形、复数、立体几何、概率、统计),见 references/math-hs2-worked-examples.md;选择性必修 14 道(数列、导数、空间向量求角、解析几何、计数原理、概率统计),见 references/math-hs3-worked-examples.md。每道题标出最容易卡的子类型,关键数值由 CI 验算;同类题生成前按 shared/ai-item-check.md 自检。
学生的错题交给通用错题本后,子类型定位见数学错误DNA 维度表第十、十一节;应用题列不出式子(数列模型、概率统计的实际问题)转应用题数学建模教练,概念本身没懂转数学概念解释器。
references/math-hs1-worked-examples.md — 高一衔接段样板题(预备知识与函数,含 CI 验算断言)references/math-hs2-worked-examples.md — 高中必修后半样板题(几何与代数、概率与统计,含 CI 验算断言)references/math-hs3-worked-examples.md — (高中)选择性必修样板题(数列、导数、空间向量、解析几何、计数原理、概率统计,含 CI 验算断言)references/claw-templates-extended.md — 5类意图分支的扩展话术和变体场景references/math-socrates-guide.md — 苏格拉底五问链各学段适配指南references/photo-4step-statemachine.md — 四步拍照法状态机定义(含意图分支、轮次预算、断点恢复与降级流程)💡 小智说:
"你发给我一道题,我最不想做的事就是直接告诉你答案。
因为那对你没有用——你今天会了这道,
明天换个数字还是不会。
我想帮你找到的,是你在哪里卡住了,
那个'卡住的地方'才是真正值得我们花时间的地方。"
© qizhitang, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
SKILL.md and 14 other files (references) in student/math/xiaozhi-math-problem-solving-coach of qizhitang/xiaozhi-skills.
Open the folder on GitHubat commit 3912b8f
Xiaozhi Math Problem Solving Coach next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Xiaozhi Math Problem Solving Coach this skillqizhitang/xiaozhi-skills | 105 | — | ~3.1k | Automated safety check: Pass | MIT | |
| DeepTutor CLIHKUDS/DeepTutor | 41k | — | ~2.8k | Automated safety check: Pass | Apache-2.0 | |
| AI Engineering Project Tutorrohitg00/ai-engineering-from-scratch | 66k | — | ~1.6k | Automated safety check: Pass | MIT | |
| Hung-Yi Lee Teaching Stylevoidful/hung-yi-lee-skill | 1.3k | — | ~13k | Automated safety check: Pass | None | |
| Claude Certification Tutorrohitg00/ai-engineering-from-scratch | 66k | — | ~3k | Automated safety check: Pass | MIT | |
| StudyVault Quiz Tutorbevibing/tutor-skills | 1.3k | — | ~1.4k | Automated safety check: Pass | MIT |
HKUDS/DeepTutor
Teaches the agent to set up and run DeepTutor from the command line: chat and capabilities, knowledge bases, partners, memory, sessions, notebooks and the server or Web app.
rohitg00/ai-engineering-from-scratch
Tutors a learner through one stage of a hands-on AI engineering project per session: lesson, prediction, code, grader run and reflection, with hints but never full solutions.
voidful/hung-yi-lee-skill
Explains machine learning, LLMs, AI agents and speech modeling in a Hung-Yi Lee-inspired teaching style, drawing on a knowledge base built from his lectures and research references.
rohitg00/ai-engineering-from-scratch
Guides a learner through one of four independent Claude certification tracks with onboarding, lessons, practice labs, mock exams and remediation.
bevibing/tutor-skills
Quizzes you on the notes in an Obsidian StudyVault, tracks proficiency per concept and drills weak areas in four-question rounds.
THU-MAIC/OpenMAIC
Designs a review-and-practice lesson around an independent first attempt, targeted feedback, supported practice, a fresh independent check and a next step.
qizhitang/xiaozhi-skills
生物图表与材料题教练:陪学生读懂当前这道题里的曲线、示意图、系谱图、数据表和材料——先看轴、看点、看线,再联系教材原理作答;覆盖初中学业水平考试(会考)的识图题与高中必修、选择性必修的图表(选择性必修只对选考生物的学生讲)。
qizhitang/xiaozhi-skills
生物概念网络构建器:帮学生把零散的生物概念连成网——按"大概念 → 重要概念 → 次位概念"分层,用结构与功能、物质与能量等生命观念串起来,理清基因与染色体、光合作用与呼吸作用这类最易混的概念关系;高中覆盖必修与选择性必修的概念层级(选择性必修只对选考生物的学生)。
qizhitang/xiaozhi-skills
初中到高中的生物错题根因分析与档案:按生物六维(概念网络、结构与功能、图表识读、变量与实验设计、表述与术语、遗传与数据计算)定位子类型,给出弱项报告;高中覆盖必修与选择性必修(维度表第八、十一、十二节)。
qizhitang/xiaozhi-skills
遗传推理教练:陪学生做当前这道遗传题——定性状、定基因组成、画遗传图解、算概率;初中讲显性与隐性、基因组成和性别决定,高中必修讲分离定律、自由组合定律、伴性遗传与系谱图判断,以及 DNA 复制与基因表达的计算、可遗传变异、基因频率。
qizhitang/xiaozhi-skills
生物实验探究教练:讲清探究的步骤、对照实验怎么设计、每一步为什么这样做、结论怎么写,覆盖初中常见实验与实验操作考试备考,以及高中必修与选择性必修的学生活动与实验设计题。
qizhitang/xiaozhi-skills
初高衔接规划:中考结束到高一上学期,诊断语文、数学、物理、化学、英语、生物各自的衔接缺口,排出暑假与高一开学头几周的衔接计划,再把具体内容交给对应的学科技能。
Categories
初中到高中的数学单题解题教练(高中覆盖必修与选择性必修):学生发来一道数学题说"卡住了""这道数学题我做错了""我不知道怎么列式"时,用追问帮学生找回思路,提示按 shared/hint-ladder.md 逐级升。. Xiaozhi Math Problem Solving Coach is an agent skill from qizhitang/xiaozhi-skills.
Xiaozhi Math Problem Solving Coach fits situations like: tasks that involve Tutoring and explanations.
Run `npx skills add qizhitang/xiaozhi-skills --skill xiaozhi-math-problem-solving-coach -a claude-code`. Or copy the skill folder (student/math/xiaozhi-math-problem-solving-coach in qizhitang/xiaozhi-skills) into .claude/skills/xiaozhi-math-problem-solving-coach in your project. Claude Code loads it when a task matches its description.
Run `npx skills add qizhitang/xiaozhi-skills --skill xiaozhi-math-problem-solving-coach -a codex`. Or copy the skill folder (student/math/xiaozhi-math-problem-solving-coach in qizhitang/xiaozhi-skills) into .agents/skills/xiaozhi-math-problem-solving-coach in your project. Codex loads it when a task matches its description.
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add qizhitang/xiaozhi-skills --skill xiaozhi-math-problem-solving-coach -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/xiaozhi-math-problem-solving-coach, .gemini/skills/xiaozhi-math-problem-solving-coach, .github/skills/xiaozhi-math-problem-solving-coach and .opencode/skills/xiaozhi-math-problem-solving-coach in your project.
SKILL.md names no scripts, command-line tools or credentials: Xiaozhi Math Problem Solving Coach is instructions for the agent only. Compatibility (from SKILL.md): WorkBuddy / SkillHub / OpenClaw / ClawHub.
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
Xiaozhi Math Problem Solving Coach is published under the MIT licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.
About 3.1k tokens (SKILL.md is roughly 12k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 19k tokens, read only when the agent opens those files.
Skills that share tags, products or a category with Xiaozhi Math Problem Solving Coach: DeepTutor CLI (HKUDS/DeepTutor, 41k stars), AI Engineering Project Tutor (rohitg00/ai-engineering-from-scratch, 66k stars), Hung-Yi Lee Teaching Style (voidful/hung-yi-lee-skill, 1.3k stars) and Claude Certification Tutor (rohitg00/ai-engineering-from-scratch, 66k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
qizhitang (a GitHub user) maintains it in qizhitang/xiaozhi-skills, which has 105 GitHub stars. The repository holds 84 skills in this directory. The repository was last updated on October 8, 2026.
Source: qizhitang/xiaozhi-skills on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.