← 返回 Skills 市场
Deepaistudy Errors
作者
stoneyshum
· GitHub ↗
· v1.0.4
· MIT-0
158
总下载
0
收藏
0
当前安装
5
版本数
在 OpenClaw 中安装
/install deepaistudy-errors
功能描述
拍照上传错题，AI自动分析生成深度学习内容与变式题，支持掌握状态标记和PDF导出，提升错题复习效率。
使用说明 (SKILL.md)
deepaistudy-errors Skill
深智智错题本 Skill。拍照上传错题，AI 自动分析并生成变式题。支持标记掌握状态。
与 deepaistudy-prep 共用同一套认证配置，无需重复登录。
工作原理
错题图片 → [AI 分析] → 错题记录
前置要求
- 已配置认证（与 deepaistudy-prep 共用）：
deepaistudy-errors config set server https://www.deepaistudy.com deepaistudy-errors config set username [email protected] deepaistudy-errors config set password your_password - AI 分析约需 1-3 分钟：请耐心等待
使用方式
添加错题（推荐）
拍照错题，上传后自动 AI 分析：
deepaistudy-errors add /path/to/error1.jpg /path/to/error2.jpg \
--subject 数学 \
--title "分数运算易错题"
会自动轮询分析状态，完成后显示结果。
列出错题
# 全部
deepaistudy-errors list
# 按学科筛选
deepaistudy-errors list --subject 数学
# 搜索关键词
deepaistudy-errors list --search "分数"
轮询分析状态
deepaistudy-errors status \x3Ctask_id> --interval 5
标记掌握状态
deepaistudy-errors master 123 # 标记为已掌握
deepaistudy-errors unmaster 123 # 取消已掌握
生成变式题
# 默认3道中等难度
deepaistudy-errors variation 123
# 指定数量和难度
deepaistudy-errors variation 123 --count 5 --difficulty hard
删除错题
deepaistudy-errors delete 123
配置
# 与 deepaistudy-prep 共用，无需单独配置
deepaistudy-errors config list
返回内容
添加成功后返回：
id：错题记录 IDsubject：学科title：标题content：错题内容（学生答案 / 正确答案 / 解析）mastery_status：掌握状态（已掌握 / 未掌握）exercise_images：原始错题图片
状态说明
| 状态 | 说明 |
|---|---|
| 已掌握 | 错题已复习完成 |
| 未掌握 | 还需要复习 |
常用命令汇总
| 命令 | 说明 |
|---|---|
add \x3Cimages> --subject X |
添加错题 + AI 分析 |
list --subject X |
错题列表 |
status \x3Ctask_id> |
轮询分析状态 |
master \x3Cnote_id> |
标记已掌握 |
unmaster \x3Cnote_id> |
取消已掌握 |
variation \x3Cnote_id> -n 3 |
生成变式题 |
delete \x3Cnote_id> |
删除错题 |
config set key value |
设置配置 |
注意事项
- 图片会压缩到最大 2048px
- 支持一次识别多道错题（自动从整页作业/试卷中提取）
- 标记已掌握后，错题不会出现在"待复习"列表
- 变式题用于举一反三，巩固同类题型
安全使用建议
This skill runs a local CLI that uploads images and the provided username/password to a server you configure. The package contains runnable Python code and a console entry point (so it will execute on installation/run). Before installing or providing credentials: 1) confirm you trust the server URL you configure (default docs mention https://www.deepaistudy.com); 2) understand your credentials will be stored in plaintext under ~/.config/deepaistudy-prep/config.ini; 3) only upload images you are comfortable sending to that server. The only minor inconsistency: the registry listed 'no install spec / instruction-only' but the package includes executable code and a setup.py — expect the platform to install/run that code rather than this being purely a docs-only skill.
能力评估
Purpose & Capability
Name/description (upload photos of wrong problems, AI analyze, generate variations) match the code and CLI commands. The code implements endpoints and commands described in SKILL.md (add/list/status/variation/master/unmaster/delete/export). The shared config location (~/.config/deepaistudy-prep) is consistent with the SKILL.md's statement about sharing auth with deepaistudy-prep.
Instruction Scope
SKILL.md instructs only to set a server, username, and password and to upload image files. The runtime code performs exactly those actions (reads images supplied by user, reads/writes config under ~/.config/deepaistudy-prep, posts to configured server endpoints, polls status). There are no instructions or code paths that read unrelated system files, access unrelated credentials, or transmit data to unexpected endpoints (server URL is configurable; default suggested domain appears in docs but the code uses the configured server).
Install Mechanism
The registry metadata stated 'No install spec — instruction-only skill', but the package includes runnable Python code, a setup.py, and a console entry point (deepaistudy-errors). That is an inconsistency to be aware of: this skill contains executable code (requests dependency) which will be installed/run by the agent platform rather than being pure prose. The code itself does not download arbitrary third-party installers or external archives.
Credentials
No environment variables or external credentials are requested by the skill metadata. The skill requires a username, password, and server URL (stored in a local config file) which are appropriate and proportionate for a service client. Note: credentials are stored in plaintext in ~/.config/deepaistudy-prep/config.ini and are sent to the configured server for authentication; users should only supply credentials for servers they trust.
Persistence & Privilege
The skill does not request 'always: true' or any elevated platform privileges. It writes only to its own config directory under the user home (~/.config/deepaistudy-prep) and does not modify other skills or system-wide agent settings.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install deepaistudy-errors - 安装完成后，直接呼叫该 Skill 的名称或使用
/deepaistudy-errors触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.4
- Documentation simplified: SKILL.md rewritten for clarity and brevity.
- Deep learning SVG analysis and PDF export descriptions removed from user guide.
- Added note on recognizing multiple questions per image (auto extraction).
- Updated AI analysis time estimate in prerequisites.
- Usage instructions, command summaries, and return values streamlined.
v1.0.3
API 切换：AI 分析接口从 error-book/ai-analyze 迁移到 breakthrough/ingest-analyze
v1.0.2
修复多图上传bug：dict key覆盖导致只发最后一张图；文件句柄关闭逻辑修复
v1.0.1
支持整页作业/试卷识别多题，每道错题独立生成一条记录；支持--preview/--count参数；支持multipart文件上传
v1.0.0
拍照添加错题，AI自动分析，共用deepaistudy-prep认证体系
元数据
常见问题
Deepaistudy Errors 是什么？
拍照上传错题，AI自动分析生成深度学习内容与变式题，支持掌握状态标记和PDF导出，提升错题复习效率。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 158 次。
如何安装 Deepaistudy Errors？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install deepaistudy-errors」即可一键安装，无需额外配置。
Deepaistudy Errors 是免费的吗？
是的，Deepaistudy Errors 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Deepaistudy Errors 支持哪些平台？
Deepaistudy Errors 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Deepaistudy Errors？
由 stoneyshum（@stoneyshum）开发并维护，当前版本 v1.0.4。
推荐 Skills