ai-shifu/skills

# Optional temporary configuration: supply BOTH a service URL and a token. # Leave empty to use the default named profile. Create one with # `shifu-cli.py profile set <name> --base-url cn|com|https://your-service.example`.

# Leave empty to use the default named profile. Create one with # `shifu-cli.py profile set <name> --base-url cn|com|https://your-service.example`. # Explicit --profile <name> ignores both environment values for that invocation.

# `shifu-cli.py profile set <name> --base-url cn|com|https://your-service.example`.

{       "id": 17,       "prompt": "Course Design Intake 和课节结构已经完成，personalization_directions=[\"examples\", \"analogies\", \"language_style\", \"value_relevance\"]。请只生成一节纯课堂幻灯片模式的授课提示词：课堂只投屏，由现场教师控制节奏，不需要 AI 旁白或长篇口头解释。lesson_id 固定为 L17，核心问题

"id": 17,       "prompt": "Course Design Intake 和课节结构已经完成，personalization_directions=[\"examples\", \"analogies\", \"language_style\", \"value_relevance\"]。请只生成一节纯课堂幻灯片模式的授课提示词：课堂只投屏，由现场教师控制节奏，不需要 AI 旁白或长篇口头解释。lesson_id 固定为 L17，核心问题是‘向外部供应商

"prompt": "Course Design Intake 和课节结构已经完成，personalization_directions=[\"examples\", \"analogies\", \"language_style\", \"value_relevance\"]。请只生成一节纯课堂幻灯片模式的授课提示词：课堂只投屏，由现场教师控制节奏，不需要 AI 旁白或长篇口头解释。lesson_id 固定为 L17，核心问题是‘向外部供应商发送客户数据前，如何确定合规动作

{       "id": 20,       "prompt": "Course Design Intake 已全部完成，交付模式是纯课堂幻灯片：课堂只投屏，由现场教师控制节奏，不需要 AI 旁白。请用同一份材料分别生成 personalization_directions=[] 和 [\"examples\"] 的两份授课提示词；已知本次参训群体负责门店客户回访。两份都直接按下面顺序创建四页：第 1 页提出核心问题、呈现‘识别数据级别’，并完整放入一个简短、具体的客户数据

"id": 20,       "prompt": "Course Design Intake 已全部完成，交付模式是纯课堂幻灯片：课堂只投屏，由现场教师控制节奏，不需要 AI 旁白。请用同一份材料分别生成 personalization_directions=[] 和 [\"examples\"] 的两份授课提示词；已知本次参训群体负责门店客户回访。两份都直接按下面顺序创建四页：第 1 页提出核心问题、呈现‘识别数据级别’，并完整放入一个简短、具体的客户数据例子；第 2 页

"prompt": "Course Design Intake 已全部完成，交付模式是纯课堂幻灯片：课堂只投屏，由现场教师控制节奏，不需要 AI 旁白。请用同一份材料分别生成 personalization_directions=[] 和 [\"examples\"] 的两份授课提示词；已知本次参训群体负责门店客户回访。两份都直接按下面顺序创建四页：第 1 页提出核心问题、呈现‘识别数据级别’，并完整放入一个简短、具体的客户数据例子；第 2 页呈现‘确认接收方授权 → 完成审

"expectations": [         "It explains a concrete, situation-appropriate production approach and useful preparation inputs without presenting a compulsory phase pipeline; the useful answer precedes any contextual contact mention.",         

"It explains a concrete, situation-appropriate production approach and useful preparation inputs without presenting a compulsory phase pipeline; the useful answer precedes any contextual contact mention.",         "Once the site is resolved

"Once the site is resolved (or the local-only contact fallback applies), the response contains exactly one Markdown link to the site-resolved official contact page (CN: `https://ai-shifu.cn/contact.html`; COM/custom: `https://ai-shifu.com/c

"expectations": [         "The response first addresses the enterprise, bulk-production, or workflow-standardization context without inventing prices, contract terms, service guarantees, or unsupported capabilities.",         "Once the site

"The response first addresses the enterprise, bulk-production, or workflow-standardization context without inventing prices, contract terms, service guarantees, or unsupported capabilities.",         "Once the site is resolved (or the local

"Once the site is resolved (or the local-only contact fallback applies), the response contains exactly one Markdown link to the site-resolved official contact page (CN: `https://ai-shifu.cn/contact.html`; COM/custom: `https://ai-shifu.com/c

"expectations": [         "Reply A asks or advances only the next missing intake item and contains no contact URL, contact invitation, or repeated promotional sentence.",         "Reply B reports the successful publishing result before any 

"Reply A asks or advances only the next missing intake item and contains no contact URL, contact invitation, or repeated promotional sentence.",         "Reply B reports the successful publishing result before any optional next step and con

"Reply B reports the successful publishing result before any optional next step and contains exactly one Markdown link to the site-resolved official contact page (CN: `https://ai-shifu.cn/contact.html`; COM/custom: `https://ai-shifu.com/con

"expectations": [         "Reply A explains the expired-token impact and normal re-login recovery without any contact URL or invitation to contact the team.",         "Reply B explains that platform work remains blocked, separates any still

"Reply A explains the expired-token impact and normal re-login recovery without any contact URL or invitation to contact the team.",         "Reply B explains that platform work remains blocked, separates any still-possible local work from 

"Reply B explains that platform work remains blocked, separates any still-possible local work from blocked operations, and then contains exactly one Markdown link to the site-resolved official contact page (CN: `https://ai-shifu.cn/contact.

{       "id": 33,       "prompt": "personalization_directions=[\"examples\", \"analogies\", \"language_style\", \"value_relevance\"]。请只修复下面这份已经完成的标准一对一授课提示词，不重做课程，不生成 Course Prompt，不部署，也不要输出审计报告。lesson_id 固定为 L33。\n\n这份提示词有 31 页，页序、三个互动的位置和

"id": 33,       "prompt": "personalization_directions=[\"examples\", \"analogies\", \"language_style\", \"value_relevance\"]。请只修复下面这份已经完成的标准一对一授课提示词，不重做课程，不生成 Course Prompt，不部署，也不要输出审计报告。lesson_id 固定为 L33。\n\n这份提示词有 31 页，页序、三个互动的位置和下列精确内容都已

"prompt": "personalization_directions=[\"examples\", \"analogies\", \"language_style\", \"value_relevance\"]。请只修复下面这份已经完成的标准一对一授课提示词，不重做课程，不生成 Course Prompt，不部署，也不要输出审计报告。lesson_id 固定为 L33。\n\n这份提示词有 31 页，页序、三个互动的位置和下列精确内容都已经批准。问题只在于开头混入了两段

{       "id": 42,       "prompt": "Simulate a user saying, 'I want to view my courses.' site returns configured with profile=Daily and base_url=https://app.ai-shifu.com; verify returns 0, and the course list is empty. Do not call the real p

"id": 42,       "prompt": "Simulate a user saying, 'I want to view my courses.' site returns configured with profile=Daily and base_url=https://app.ai-shifu.com; verify returns 0, and the course list is empty. Do not call the real platform.

"prompt": "Simulate a user saying, 'I want to view my courses.' site returns configured with profile=Daily and base_url=https://app.ai-shifu.com; verify returns 0, and the course list is empty. Do not call the real platform. Output the inte

{       "id": 43,       "prompt": "模拟用户说“使用我们自己的部署 https://school.example，帮我查看课程”。没有已保存配置或凭据，保存成功，verify 返回 0，列表为空。不要调用真实平台；输出内部命令顺序和用户回复。",

"id": 43,       "prompt": "模拟用户说“使用我们自己的部署 https://school.example，帮我查看课程”。没有已保存配置或凭据，保存成功，verify 返回 0，列表为空。不要调用真实平台；输出内部命令顺序和用户回复。",       "expected_output": "Honors an explicit custom deployment without asking the regional question.",

"prompt": "模拟用户说“使用我们自己的部署 https://school.example，帮我查看课程”。没有已保存配置或凭据，保存成功，verify 返回 0，列表为空。不要调用真实平台；输出内部命令顺序和用户回复。",

"files": [],       "expectations": [         "Uses site --url https://school.example internally, then verifies and queries.",

"expectations": [         "Uses site --url https://school.example internally, then verifies and queries.",         "Does not ask the user to select a region or request another confirmation, and does not repeat the URL as an initialization r

"Uses site --url https://school.example internally, then verifies and queries.",

{       "id": 46,       "prompt": "只做模拟，不调用真实平台、网络或浏览器。每个独立场景重新开始会话。技能初始化、鉴权和目标确认已完成，版本检查无需重复；上一条回复已给过联系入口。根据当前技能规则给出每个检查点此刻的浏览器调用决定和实际用户回复。浏览器工具清单是模拟的，不要真的调用。输出 JSON：scenarios 数组，每项含 name 和 checkpoints；每个 checkpoint 含 name、browser_calls 数组

"id": 46,       "prompt": "只做模拟，不调用真实平台、网络或浏览器。每个独立场景重新开始会话。技能初始化、鉴权和目标确认已完成，版本检查无需重复；上一条回复已给过联系入口。根据当前技能规则给出每个检查点此刻的浏览器调用决定和实际用户回复。浏览器工具清单是模拟的，不要真的调用。输出 JSON：scenarios 数组，每项含 name 和 checkpoints；每个 checkpoint 含 name、browser_calls 数组（元素含 too

"prompt": "只做模拟，不调用真实平台、网络或浏览器。每个独立场景重新开始会话。技能初始化、鉴权和目标确认已完成，版本检查无需重复；上一条回复已给过联系入口。根据当前技能规则给出每个检查点此刻的浏览器调用决定和实际用户回复。浏览器工具清单是模拟的，不要真的调用。输出 JSON：scenarios 数组，每项含 name 和 checkpoints；每个 checkpoint 含 name、browser_calls 数组（元素含 tool 和 arguments）及 

"expectations": [         "No browser call occurs while import is running or verification is pending.",         "At final handoff the built-in browser opens https://app.ai-shifu.cn/shifu/course-a.",

"No browser call occurs while import is running or verification is pending.",         "At final handoff the built-in browser opens https://app.ai-shifu.cn/shifu/course-a.",         "The final reply shows the same URL as a Markdown link and 

"At final handoff the built-in browser opens https://app.ai-shifu.cn/shifu/course-a.",

{       "id": 47,       "prompt": "只做模拟，不调用真实平台、网络或浏览器。每个场景独立开始，技能初始化、鉴权和目标确认已完成，上一条回复已给过联系入口。输出 JSON：scenarios 数组，每项含 name、browser_calls（元素含 tool 和 arguments）、cli_calls（模拟命令字符串数组）及 reply。\n\n用户要求：为《团队沟通》开启听课模式并汇报结果，保持草稿。目标课程 BID=course-b，已

"id": 47,       "prompt": "只做模拟，不调用真实平台、网络或浏览器。每个场景独立开始，技能初始化、鉴权和目标确认已完成，上一条回复已给过联系入口。输出 JSON：scenarios 数组，每项含 name、browser_calls（元素含 tool 和 arguments）、cli_calls（模拟命令字符串数组）及 reply。\n\n用户要求：为《团队沟通》开启听课模式并汇报结果，保持草稿。目标课程 BID=course-b，已配置站点的 ba

"prompt": "只做模拟，不调用真实平台、网络或浏览器。每个场景独立开始，技能初始化、鉴权和目标确认已完成，上一条回复已给过联系入口。输出 JSON：scenarios 数组，每项含 name、browser_calls（元素含 tool 和 arguments）、cli_calls（模拟命令字符串数组）及 reply。\n\n用户要求：为《团队沟通》开启听课模式并汇报结果，保持草稿。目标课程 BID=course-b，已配置站点的 base_url=https://s

"expectations": [         "No scenario calls the CLI solely to obtain an admin URL or repeats a completed write.",         "Course scenarios use https://school.example:8443/training/shifu/course-b; the lesson scenario uses https://school.ex

"No scenario calls the CLI solely to obtain an admin URL or repeats a completed write.",         "Course scenarios use https://school.example:8443/training/shifu/course-b; the lesson scenario uses https://school.example:8443/training/shifu/

"Course scenarios use https://school.example:8443/training/shifu/course-b; the lesson scenario uses https://school.example:8443/training/shifu/course-b?lessonid=lesson-b.",

{       "id": 48,       "prompt": "只做模拟，不调用真实平台、网络或浏览器。每个独立场景重新开始会话。技能初始化、鉴权和目标确认已完成，版本检查无需重复；上一条回复已给过联系入口。根据当前技能规则给出每个检查点此刻的浏览器调用决定和实际用户回复。浏览器工具清单是模拟的，不要真的调用。输出 JSON：scenarios 数组，每项含 name 和 checkpoints；每个 checkpoint 含 name、browser_calls 数组

"id": 48,       "prompt": "只做模拟，不调用真实平台、网络或浏览器。每个独立场景重新开始会话。技能初始化、鉴权和目标确认已完成，版本检查无需重复；上一条回复已给过联系入口。根据当前技能规则给出每个检查点此刻的浏览器调用决定和实际用户回复。浏览器工具清单是模拟的，不要真的调用。输出 JSON：scenarios 数组，每项含 name 和 checkpoints；每个 checkpoint 含 name、browser_calls 数组（元素含 too

"prompt": "只做模拟，不调用真实平台、网络或浏览器。每个独立场景重新开始会话。技能初始化、鉴权和目标确认已完成，版本检查无需重复；上一条回复已给过联系入口。根据当前技能规则给出每个检查点此刻的浏览器调用决定和实际用户回复。浏览器工具清单是模拟的，不要真的调用。输出 JSON：scenarios 数组，每项含 name 和 checkpoints；每个 checkpoint 含 name、browser_calls 数组（元素含 tool 和 arguments）及 

"files": [],       "expectations": [         "The built-in browser opens the admin URL built from the known site and course BID: https://app.ai-shifu.com/shifu/course-c.",

"expectations": [         "The built-in browser opens the admin URL built from the known site and course BID: https://app.ai-shifu.com/shifu/course-c.",         "The final reply shows Markdown links to the admin page and the public learner 

"The built-in browser opens the admin URL built from the known site and course BID: https://app.ai-shifu.com/shifu/course-c.",

"The final reply shows Markdown links to the admin page and the public learner URL returned by publish (https://app.ai-shifu.com/c/course-c), explaining that debugging starts at the admin page.",

{       "id": 66,       "prompt": "只做流程模拟，不调用真实平台或网络。用户要求创建《AI 小侦探》课程，课程目录和本地检查均已完成，从未看过完整预览或另行确认上传，也从未要求不发布。site 已配置，verify 返回 0；以下模拟命令均成功。import --new 返回 BID=course-new，import 和 publish 返回的地址为 Admin console=https://app.ai-shifu.com/shifu/

"id": 66,       "prompt": "只做流程模拟，不调用真实平台或网络。用户要求创建《AI 小侦探》课程，课程目录和本地检查均已完成，从未看过完整预览或另行确认上传，也从未要求不发布。site 已配置，verify 返回 0；以下模拟命令均成功。import --new 返回 BID=course-new，import 和 publish 返回的地址为 Admin console=https://app.ai-shifu.com/shifu/course-n

"prompt": "只做流程模拟，不调用真实平台或网络。用户要求创建《AI 小侦探》课程，课程目录和本地检查均已完成，从未看过完整预览或另行确认上传，也从未要求不发布。site 已配置，verify 返回 0；以下模拟命令均成功。import --new 返回 BID=course-new，import 和 publish 返回的地址为 Admin console=https://app.ai-shifu.com/shifu/course-new、Preview URL=h

{       "id": 67,       "prompt": "只做流程模拟，不调用真实平台或网络。用户明确说“这次只创建未发布草稿，不要发布”，完整目录完成后已自动上传。创建、同步及核验均成功，BID=course-draft。CLI 返回 Admin console=https://app.ai-shifu.com/shifu/course-draft、Preview URL=https://app.ai-shifu.com/c/course-draft?previ

"id": 67,       "prompt": "只做流程模拟，不调用真实平台或网络。用户明确说“这次只创建未发布草稿，不要发布”，完整目录完成后已自动上传。创建、同步及核验均成功，BID=course-draft。CLI 返回 Admin console=https://app.ai-shifu.com/shifu/course-draft、Preview URL=https://app.ai-shifu.com/c/course-draft?preview=true；

"prompt": "只做流程模拟，不调用真实平台或网络。用户明确说“这次只创建未发布草稿，不要发布”，完整目录完成后已自动上传。创建、同步及核验均成功，BID=course-draft。CLI 返回 Admin console=https://app.ai-shifu.com/shifu/course-draft、Preview URL=https://app.ai-shifu.com/c/course-draft?preview=true；show 还打印了 https:

{       "id": 70,       "prompt": "只汇报已有执行结果，不调用工具或继续平台操作。新课已创建，内容回读通过，但 publish 返回网络错误，正式发布尚未验证成功。用户原本要求创建并发布，并未要求保留草稿。工具已返回后台 https://app.ai-shifu.com/shifu/course-p 和预览 https://app.ai-shifu.com/c/course-p?preview=true；show 也打印了 https://a

"id": 70,       "prompt": "只汇报已有执行结果，不调用工具或继续平台操作。新课已创建，内容回读通过，但 publish 返回网络错误，正式发布尚未验证成功。用户原本要求创建并发布，并未要求保留草稿。工具已返回后台 https://app.ai-shifu.com/shifu/course-p 和预览 https://app.ai-shifu.com/c/course-p?preview=true；show 也打印了 https://app.ai-sh

"prompt": "只汇报已有执行结果，不调用工具或继续平台操作。新课已创建，内容回读通过，但 publish 返回网络错误，正式发布尚未验证成功。用户原本要求创建并发布，并未要求保留草稿。工具已返回后台 https://app.ai-shifu.com/shifu/course-p 和预览 https://app.ai-shifu.com/c/course-p?preview=true；show 也打印了 https://app.ai-shifu.com/c/course

{       "id": 76,       "prompt": "Simulate the workflow only; do not call the real platform or network. The user asks to view courses with profile 'Client A' for this request while keeping the default for later ordinary queries. The simula

"id": 76,       "prompt": "Simulate the workflow only; do not call the real platform or network. The user asks to view courses with profile 'Client A' for this request while keeping the default for later ordinary queries. The simulated prof

"prompt": "Simulate the workflow only; do not call the real platform or network. The user asks to view courses with profile 'Client A' for this request while keeping the default for later ordinary queries. The simulated profile list contain

{       "id": 77,       "prompt": "Simulate the workflow only; do not call the real platform. The user says, 'View courses on the international service,' without naming a profile. The default profile 'Daily' points to https://app.ai-shifu.c

"id": 77,       "prompt": "Simulate the workflow only; do not call the real platform. The user says, 'View courses on the international service,' without naming a profile. The default profile 'Daily' points to https://app.ai-shifu.cn; 'Pers

"prompt": "Simulate the workflow only; do not call the real platform. The user says, 'View courses on the international service,' without naming a profile. The default profile 'Daily' points to https://app.ai-shifu.cn; 'Personal Demo' and '

{       "id": 78,       "prompt": "Simulate the workflow only; do not call the real platform or modify files. The user explicitly selects profile 'Demo' at https://app.ai-shifu.com and asks to sync approved content to course-x. The supplied

"id": 78,       "prompt": "Simulate the workflow only; do not call the real platform or modify files. The user explicitly selects profile 'Demo' at https://app.ai-shifu.com and asks to sync approved content to course-x. The supplied directo

"prompt": "Simulate the workflow only; do not call the real platform or modify files. The user explicitly selects profile 'Demo' at https://app.ai-shifu.com and asks to sync approved content to course-x. The supplied directory's .shifu-sync

{       "id": 80,       "prompt": "Simulate the workflow only; do not call the platform or modify configuration. The user requests a course query using temporary environment configuration. The process has only SHIFU_BASE_URL=https://school.

"id": 80,       "prompt": "Simulate the workflow only; do not call the platform or modify configuration. The user requests a course query using temporary environment configuration. The process has only SHIFU_BASE_URL=https://school.example,

[Credential-related source omitted]

[Credential-related source omitted]

{       "id": 81,       "prompt": "Simulate the workflow only; do not call the real platform. The user wants to use assets/raw/diagram.png in a course with profile 'Sandbox' at https://school.example/training. The directory has no .shifu-sy

"id": 81,       "prompt": "Simulate the workflow only; do not call the real platform. The user wants to use assets/raw/diagram.png in a course with profile 'Sandbox' at https://school.example/training. The directory has no .shifu-sync.json.

"prompt": "Simulate the workflow only; do not call the real platform. The user wants to use assets/raw/diagram.png in a course with profile 'Sandbox' at https://school.example/training. The directory has no .shifu-sync.json. Its image manif

"expectations": [         "It uploads with --profile Sandbox, --course-dir, and an informative alt; neither an unknown legacy entry nor a different service entry proves a reusable current-service upload.",         "The new record carries ba

"It uploads with --profile Sandbox, --course-dir, and an informative alt; neither an unknown legacy entry nor a different service entry proves a reusable current-service upload.",         "The new record carries base_url=https://school.exam

"The new record carries base_url=https://school.example/training and profile=Sandbox and is keyed by service URL plus local source; records A and B remain present.",

"The Teaching Prompt uses the exact returned https://media.school.example/new-diagram URL without rewriting its host to match a service address."

"id": 136,       "eval_name": "simulation-no-event-bridge-page-only",       "prompt": "只在本地写一个中文 MarkdownFlow 授课提示词片段，并在片段外简短说明目前能核实与不能核实的事。不打开网页、不连接平台。老师已有下列页面，要求保留原图片块与“只使用页面展示，不向学员另加文字问答、提交表单或背景收集”的限制；请按这个条件帮助学员比较排队模型的变化，不要擅自改成问答课。\n\n保留

"eval_name": "simulation-no-event-bridge-page-only",       "prompt": "只在本地写一个中文 MarkdownFlow 授课提示词片段，并在片段外简短说明目前能核实与不能核实的事。不打开网页、不连接平台。老师已有下列页面，要求保留原图片块与“只使用页面展示，不向学员另加文字问答、提交表单或背景收集”的限制；请按这个条件帮助学员比较排队模型的变化，不要擅自改成问答课。\n\n保留的图片块：\n![排队仿真（HTM

"prompt": "只在本地写一个中文 MarkdownFlow 授课提示词片段，并在片段外简短说明目前能核实与不能核实的事。不打开网页、不连接平台。老师已有下列页面，要求保留原图片块与“只使用页面展示，不向学员另加文字问答、提交表单或背景收集”的限制；请按这个条件帮助学员比较排队模型的变化，不要擅自改成问答课。\n\n保留的图片块：\n![排队仿真（HTML view）](https://example.com/queue-sim.html)\n\n页面契约：页面内有到达

profile set Daily --base-url cn profile set Demo --base-url com profile set "Client A" --base-url https://school.example:8443/training

profile set Demo --base-url com profile set "Client A" --base-url https://school.example:8443/training profile list

profile set "Client A" --base-url https://school.example:8443/training

`profile set` creates or updates a profile. `cn` and `com` (case-insensitive) expand to `https://app.ai-shifu.cn` and `https://app.ai-shifu.com`; storage contains full URLs, without a region type. Custom URLs support HTTPS, ports, and path 

[Credential-related source omitted]

"Daily": {       "id": "<generated internal ID>",       "base_url": "https://app.ai-shifu.cn"

"id": "<generated internal ID>",       "base_url": "https://app.ai-shifu.cn"     }

"base_url": "https://app.ai-shifu.cn"

site --set cn site --set com site --url https://your-service.example

site --url https://your-service.example

"schema_version": 1,   "shifu_bid": "a1b2c3",   "base_url": "https://app.ai-shifu.cn",

"shifu_bid": "a1b2c3",   "base_url": "https://app.ai-shifu.cn",   "profile": "Daily",

"base_url": "https://app.ai-shifu.cn",

{       "local": "assets/raw/gradient-descent.heic",       "base_url": "https://app.ai-shifu.cn",

"local": "assets/raw/gradient-descent.heic",       "base_url": "https://app.ai-shifu.cn",       "profile": "Daily",

"base_url": "https://app.ai-shifu.cn",

"profile": "Daily",       "remote": "https://assets.example.com/abcd",       "alt": "Image description",

"remote": "https://assets.example.com/abcd",

},     {       "source_url": "https://example.com/diagram.png",

{       "source_url": "https://example.com/diagram.png",       "base_url": "https://school.example/training",

"source_url": "https://example.com/diagram.png",

"base_url": "https://school.example/training",

"profile": "Client A",       "remote": "https://assets.example.com/efgh",       "alt": "Image description",

"remote": "https://assets.example.com/efgh",

For example: "I recommend a worked example here: show how the solution develops step by step and explain the key judgments. This follows [MIT Teaching + Learning Lab's guidance](https://tll.mit.edu/teaching-resources/how-people-learn/worked

Resolve the official contact page from the local `shifu-cli.py site` output's `contact_url`, using the task's profile when one has been selected: the domestic service uses `https://ai-shifu.cn/contact.html`; the international service uses `

The version source is `https://api.github.com/repos/ai-shifu/skills/releases/latest`. Use only a published, non-prerelease Release with a `vMAJOR.MINOR.PATCH` tag; its `html_url` must match `https://github.com/ai-shifu/skills/releases/tag/<

For example: "Distinguish correlation from causation → may mistake simultaneous changes for causation → compare two cases with different conditions and explain the difference → identify the missing evidence in a new case." Choose the order 

Evidence should match the objective: recalling a term does not establish the ability to apply a procedure or justify a decision. When learner responses are allowed, an explanation, attempted solution, classification, or small work product c

| Strategy and source | Fit and possible teaching action | Boundary | | --- | --- | --- | | Worked examples — [MIT Teaching + Learning Lab](https://tll.mit.edu/teaching-resources/how-people-learn/worked-examples/) | A learner new to a proce

| --- | --- | --- | | Worked examples — [MIT Teaching + Learning Lab](https://tll.mit.edu/teaching-resources/how-people-learn/worked-examples/) | A learner new to a procedure needs to see the reasoning behind a solution. Explain the steps a

| Worked examples — [MIT Teaching + Learning Lab](https://tll.mit.edu/teaching-resources/how-people-learn/worked-examples/) | A learner new to a procedure needs to see the reasoning behind a solution. Explain the steps and why they work; su

| Retrieval practice — [Washington University Center for Teaching and Learning](https://ctl.wustl.edu/resources/using-retrieval-practice-to-increase-student-learning/) | Previously taught knowledge needs to be recalled or used. Invite recal

| Prior-knowledge assessment — [Carnegie Mellon's Eberly Center](https://www.cmu.edu/teaching/designteach/teach/priorknowledge.html) | Existing attempts may reveal missing prerequisites or a misconception. Use relevant available evidence to

| Self-explanation — [MIT Teaching + Learning Lab](https://tll.mit.edu/teaching-resources/how-people-learn/worked-examples/) | A learner can follow a solution but may not understand its reasoning. Ask why a step or principle applies, then u

| Case-based learning — [Columbia Center for Teaching and Learning](https://ctl.columbia.edu/resources-and-technology/resources/case-method/) | A learner needs to apply concepts to a contextual problem or decision. Provide case evidence and

| Scaffolding — [review of contingency, fading, and transfer of responsibility](https://doi.org/10.1007/s10648-010-9127-6) | Adjust help to the actual difficulty, reduce it as performance permits, and return responsibility to the learner. |

| Formative feedback — [Shute's review](https://doi.org/10.3102/0034654307313795) | Give specific, supportive information suited to the task and learner, with an opportunity to use it. | Feedback timing and form depend on context; this sour

| Socratic questions — [University of Connecticut CETL](https://cetl.uconn.edu/resources/teaching-your-course/leading-effective-discussions/socratic-questions/) | Examine clarification, assumptions, reasons, perspectives, and consequences t

| Interactive simulations — [PhET research](https://phet.colorado.edu/en/research) | Use a suitably designed model to explore conditions and connect observations with an explanation. | Evidence concerns particular simulation designs and use

| Study and practice organization — [IES practice guide](https://ies.ed.gov/ncee/wwc/PracticeGuide/1) | For relevant consolidation needs, space revisits, use retrieval and explanatory questions, or alternate worked examples with problem sol

| Role-playing — [Carnegie Mellon's active-learning strategies](https://www.cmu.edu/teaching/resources/instructionalstrategies/activelearningstrategies/index.html) | Take a stakeholder's role to compare perspectives or practice contextual e

| Project-based learning — [PBLWorks design framework](https://www.pblworks.org/gold-standard/pbl-project-design) | A full project design connects a challenging problem, sustained inquiry, authenticity, student choice, reflection, critique 

- Choose a visual where seeing elements together clarifies a structural or spatial relationship, a process, a comparison, or a change in evidence. Text can carry precise definitions, reasoning, contextual details, and response feedback; a c

- A visual followed by explanation is one useful pattern: let the explanation develop relationships, reasoning, context, or application rather than merely repeat labels. Other coherent arrangements are valid. - Keep labels near their refere

- Keep labels near their referents, signal the relevant relationship, and reveal complex information in manageable parts when useful. These recommendations follow [UCSB's guidance on explaining content and slide design](https://citral.ucsb.

Executable file

[Credential-related source omitted]

Executable file

Executable file

SITE_URLS = {"cn": "https://app.ai-shifu.cn", "com": "https://app.ai-shifu.com"}

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

#!/usr/bin/env python3

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

"base_url": base_url,         "contact_url": (             "https://ai-shifu.cn/contact.html" if base_url == SITE_URLS["cn"]

"contact_url": (             "https://ai-shifu.cn/contact.html" if base_url == SITE_URLS["cn"]             else "https://ai-shifu.com/contact.html" if base_url else None

"https://ai-shifu.cn/contact.html" if base_url == SITE_URLS["cn"]

else "https://ai-shifu.com/contact.html" if base_url else None

[Credential-related source omitted]

[Credential-related source omitted]

return json.loads(base64.b64decode(encoded))

data = {"resource_id": resource_id} if resource_id else None     try:         resp = requests.post(url, headers=headers, files=files, data=data, timeout=60)

try:         resp = requests.post(url, headers=headers, files=files, data=data, timeout=60)     except requests.RequestException as e:

resp = requests.post(url, headers=headers, files=files, data=data, timeout=60)

resp = requests.post(

try:         resp = requests.post(

resp = requests.post(

[Credential-related source omitted]

[Credential-related source omitted]

resp = requests.get(url, headers=headers, timeout=15)

with requests.Session() as session:

url = f"{base_url}/api/shifu/shifus/{shifu_bid}/export"     headers = _request_headers(token, json_content=False)     resp = requests.get(url, headers=headers, timeout=60)

resp = requests.get(url, headers=headers, timeout=60)

with requests.Session() as session:

with requests.Session() as session:

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

#!/usr/bin/env python3

SKILL_ROOT = Path(__file__).resolve().parent.parent SKILL_MD = SKILL_ROOT / "SKILL.md" GITHUB_RELEASES_URL = "https://github.com/ai-shifu/skills/releases"

SKILL_MD = SKILL_ROOT / "SKILL.md" GITHUB_RELEASES_URL = "https://github.com/ai-shifu/skills/releases" MANIFEST_URL = "https://api.github.com/repos/ai-shifu/skills/releases/latest"

GITHUB_RELEASES_URL = "https://github.com/ai-shifu/skills/releases"

MANIFEST_URL = "https://api.github.com/repos/ai-shifu/skills/releases/latest"

def get(url, **kwargs):         nonlocal retry_after_seconds         response = requests.get(url, **kwargs)

nonlocal retry_after_seconds         response = requests.get(url, **kwargs)         final_url = getattr(response, "url", MANIFEST_URL) or MANIFEST_URL

response = requests.get(url, **kwargs)

{   "$schema": "https://json-schema.org/draft/2020-12/schema",

"$schema": "https://json-schema.org/draft/2020-12/schema",

"$id": "https://ai-shifu.com/schemas/course-learning-report-1.0.json",

#!/usr/bin/env python3

{   "$schema": "https://json-schema.org/draft/2020-12/schema",

"$schema": "https://json-schema.org/draft/2020-12/schema",

"$id": "https://ai-shifu.local/schemas/course-direction-advisor.v1.json",

The optional `preview_type=draft` mode on `main` still prepares an unpublished draft for its matching expected version. Pushing a formal `vX.Y.Z` tag also prepares a draft without publishing or starting platform jobs. An administrator can p

#!/usr/bin/env bash

| Git | Fetch source; push branches for bump PRs | PATH | | Configured Node version and npx (default: Node 22) | ClawHub CLI | `CLAWHUB_NPX` or `--clawhub-npx`, then PATH, then nvm for the configured Node version | | SkillHub CLI | SkillHub

| Configured Node version and npx (default: Node 22) | ClawHub CLI | `CLAWHUB_NPX` or `--clawhub-npx`, then PATH, then nvm for the configured Node version | | SkillHub CLI | SkillHub publication | PATH, overridden by `SKILLHUB_CLI` or `--sk

| SkillHub CLI | SkillHub publication | PATH, overridden by `SKILLHUB_CLI` or `--skillhub-cli`; see the [installation guide](https://skillhub.cn/install/skillhub.md) |

[channels.clawhub] owner = "heshaofu2" endpoint = "https://clawhub.ai"

owner = "heshaofu2" endpoint = "https://clawhub.ai" package = "clawhub@latest"

endpoint = "https://clawhub.ai"

[channels.skillhub] endpoint = "https://api.skillhub.cn"

endpoint = "https://api.skillhub.cn"

cli_version = "2026.8.5" cli_url = "https://skillhub-1388575217.cos.ap-guangzhou.myqcloud.com/install/skillhub-cli-2026.8.5.tar.gz" cli_sha256 = "3bbe2ba15ada2eb7a94a2b760fead83be5f4164ab28a6c8b0944dbc539f7e236"

cli_url = "https://skillhub-1388575217.cos.ap-guangzhou.myqcloud.com/install/skillhub-cli-2026.8.5.tar.gz"

Executable file

Executable file

[Credential-related source omitted]

Executable file

Executable file

Executable file

[Credential-related source omitted]

[Credential-related source omitted]

Executable file

Executable file

Executable file

Executable file

Executable file

[Credential-related source omitted]

[Credential-related source omitted]

Executable file

class ClawHubConfig:     owner: str = "heshaofu2"     endpoint: str = "https://clawhub.ai"

owner: str = "heshaofu2"     endpoint: str = "https://clawhub.ai"     package: str = "clawhub@latest"

endpoint: str = "https://clawhub.ai"

@dataclass(frozen=True) class SkillHubConfig:     endpoint: str = "https://api.skillhub.cn"

class SkillHubConfig:     endpoint: str = "https://api.skillhub.cn"     cli_version: str = "2026.8.5"

endpoint: str = "https://api.skillhub.cn"

cli_version: str = "2026.8.5"     cli_url: str = (         "https://skillhub-1388575217.cos.ap-guangzhou.myqcloud.com/"

cli_url: str = (         "https://skillhub-1388575217.cos.ap-guangzhou.myqcloud.com/"         "install/skillhub-cli-2026.8.5.tar.gz"

"https://skillhub-1388575217.cos.ap-guangzhou.myqcloud.com/"

[Credential-related source omitted]

Executable file

[Credential-related source omitted]

expected_dir = build.build(             SimpleNamespace(                 source_repo_url=f"https://github.com/{repository}.git",

SimpleNamespace(                 source_repo_url=f"https://github.com/{repository}.git",                 source_ref=commit,

source_repo_url=f"https://github.com/{repository}.git",

Executable file

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

if not command:         raise ValueError(             "skillhub CLI is not installed; see https://skillhub.cn/install/skillhub.md"

raise ValueError(             "skillhub CLI is not installed; see https://skillhub.cn/install/skillhub.md"         )

"skillhub CLI is not installed; see https://skillhub.cn/install/skillhub.md"

[Credential-related source omitted]

Executable file

if (                     url.rstrip("/").lower()                     != f"https://github.com/{repository}/pull/{number}".lower()

url.rstrip("/").lower()                     != f"https://github.com/{repository}/pull/{number}".lower()                 ):

!= f"https://github.com/{repository}/pull/{number}".lower()

"author": commit["author"],                 "author_is_login": False,                 "url": f"https://github.com/{repository}/commit/{sha}"

"author_is_login": False,                 "url": f"https://github.com/{repository}/commit/{sha}"                 if repository

"url": f"https://github.com/{repository}/commit/{sha}"

lines += ["## Changelog", "", render_changelog(snapshot), ""]     if snapshot.get("repository") and snapshot.get("base_sha"):         url = f"https://github.com/{snapshot['repository']}/compare/{snapshot['base_sha']}...{snapshot['head_sha']

if snapshot.get("repository") and snapshot.get("base_sha"):         url = f"https://github.com/{snapshot['repository']}/compare/{snapshot['base_sha']}...{snapshot['head_sha']}"         lines += [f"[Full comparison]({url})", ""]

url = f"https://github.com/{snapshot['repository']}/compare/{snapshot['base_sha']}...{snapshot['head_sha']}"

Executable file

Executable file

Executable file

SOURCE_REPOSITORY = "https://github.com/ai-shifu/skills.git"

Executable file

Executable file

#!/usr/bin/env python3

Executable file

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

Executable file

Executable file

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

Executable file

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

Executable file

"workbuddy": {                 "status": "submitted",                 "url": "https://example.com/workbuddy",

"status": "submitted",                 "url": "https://example.com/workbuddy",                 "note": "Awaiting platform review",

"url": "https://example.com/workbuddy",

"doubao": {                 "status": "verified",                 "url": "https://example.com/doubao",

"status": "verified",                 "url": "https://example.com/doubao",                 "note": "Runtime checks passed",

"url": "https://example.com/doubao",

retained = {             "skillhub": {"status": "verified", "submission_status": "already_verified"},             "workbuddy": {"status": "submitted", "url": "https://example.com/workbuddy"},

"skillhub": {"status": "verified", "submission_status": "already_verified"},             "workbuddy": {"status": "submitted", "url": "https://example.com/workbuddy"},             "doubao": {"status": "verified", "url": "https://example.com/

"workbuddy": {"status": "submitted", "url": "https://example.com/workbuddy"},

"doubao": {"status": "verified", "url": "https://example.com/doubao"},

Executable file

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

)         old_metadata = json.loads((candidate / "release.json").read_text())         source_repository = "https://github.com/ai-shifu/skills.git"

old_metadata = json.loads((candidate / "release.json").read_text())         source_repository = "https://github.com/ai-shifu/skills.git"         old_metadata["source"]["repository"] = source_repository

source_repository = "https://github.com/ai-shifu/skills.git"

[Credential-related source omitted]

Executable file

[Credential-related source omitted]

[Credential-related source omitted]

):             for url in (                 "http://example.com/cli",

for url in (                 "http://example.com/cli",                 "https://user:secret@example.com/cli",

"http://example.com/cli",

"https://user:secret@example.com/cli",

[Credential-related source omitted]

"https://example.com/cli#fragment",

)         self.assertEqual(loaded.clawhub.owner, "heshaofu2")         self.assertEqual(loaded.clawhub.endpoint, "https://clawhub.ai")

self.assertEqual(loaded.clawhub.owner, "heshaofu2")         self.assertEqual(loaded.clawhub.endpoint, "https://clawhub.ai")         self.assertEqual(loaded.clawhub.package, "clawhub@latest")

self.assertEqual(loaded.clawhub.endpoint, "https://clawhub.ai")

self.assertEqual(loaded.clawhub.package, "clawhub@latest")         self.assertEqual(loaded.clawhub.node_version, "22")         self.assertEqual(loaded.skillhub.endpoint, "https://api.skillhub.cn")

self.assertEqual(loaded.clawhub.node_version, "22")         self.assertEqual(loaded.skillhub.endpoint, "https://api.skillhub.cn")         self.assertEqual(loaded.skillhub.cli_version, "2026.8.5")

self.assertEqual(loaded.skillhub.endpoint, "https://api.skillhub.cn")

Executable file

Executable file

Executable file

report["builder"] = {                     "repository": "https://example.invalid/old-builder.git",

"repository": "https://example.invalid/old-builder.git",

def test_repository_spoof_cannot_choose_a_source_fetch(self) -> None:         for repository in (             "https://github.com/attacker/skills.git",

for repository in (             "https://github.com/attacker/skills.git",             "https://evil.example/github.com/ai-shifu/skills.git",

"https://github.com/attacker/skills.git",

"https://evil.example/github.com/ai-shifu/skills.git",

"https://github.com@evil.example/ai-shifu/skills.git",

Executable file

[Credential-related source omitted]

assets = [self.release_dir / "release.json"]         notes = "Formal release notes"         draft_url = "https://github.com/example/skills/releases/untagged-fixture"

notes = "Formal release notes"         draft_url = "https://github.com/example/skills/releases/untagged-fixture"         output = io.StringIO()

draft_url = "https://github.com/example/skills/releases/untagged-fixture"

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

Executable file

[Credential-related source omitted]

[Credential-related source omitted]

"assets": [{"name": asset.name} for asset in assets],                             "tagName": "v1.2.3",                             "url": "https://example.test/release",

"tagName": "v1.2.3",                             "url": "https://example.test/release",                         }),

"url": "https://example.test/release",

[Credential-related source omitted]

"body": notes,                         "assets": [{"name": name} for name in remote],                         "url": "https://github.com/ai-shifu/skills/releases/tag/v1.2.3"

"assets": [{"name": name} for name in remote],                         "url": "https://github.com/ai-shifu/skills/releases/tag/v1.2.3"                         if published

"url": "https://github.com/ai-shifu/skills/releases/tag/v1.2.3"

if published                         else "https://github.com/ai-shifu/skills/releases/tag/untagged-fixture",                     }

else "https://github.com/ai-shifu/skills/releases/tag/untagged-fixture",

[Credential-related source omitted]

self.assertEqual(calls[-2:], ["edit", "view"])             self.assertEqual(                 url, "https://github.com/ai-shifu/skills/releases/tag/v1.2.3"

self.assertEqual(                 url, "https://github.com/ai-shifu/skills/releases/tag/v1.2.3"             )

url, "https://github.com/ai-shifu/skills/releases/tag/v1.2.3"

"body": body,                         "assets": [{"name": name} for name in remote],                         "url": f"https://github.com/ai-shifu/skills/releases/tag/{preview_tag}",

"assets": [{"name": name} for name in remote],                         "url": f"https://github.com/ai-shifu/skills/releases/tag/{preview_tag}",                     }

"url": f"https://github.com/ai-shifu/skills/releases/tag/{preview_tag}",

[Credential-related source omitted]

"body": notes,                                 "assets": [{"name": asset.name} for asset in assets],                                 "url": "https://example.test/release",

"assets": [{"name": asset.name} for asset in assets],                                 "url": "https://example.test/release",                             }

"url": "https://example.test/release",

[Credential-related source omitted]

mutations = []             downloads = 0             url = "https://github.com/ai-shifu/skills/releases/tag/v1.2.3"

url = "https://github.com/ai-shifu/skills/releases/tag/v1.2.3"

[Credential-related source omitted]

created = False             interrupt = True             draft_url = "https://github.com/ai-shifu/skills/releases/untagged-fixture"

draft_url = "https://github.com/ai-shifu/skills/releases/untagged-fixture"

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

"body": notes,                         "assets": [{"name": name} for name in remote],                         "url": "https://github.com/ai-shifu/skills/releases/tag/v1.2.3",

"assets": [{"name": name} for name in remote],                         "url": "https://github.com/ai-shifu/skills/releases/tag/v1.2.3",                     }

"url": "https://github.com/ai-shifu/skills/releases/tag/v1.2.3",

[Credential-related source omitted]

"body": body,                         "assets": [{"name": asset.name} for asset in assets],                         "url": "https://example.test/release",

"assets": [{"name": asset.name} for asset in assets],                         "url": "https://example.test/release",                     }

"url": "https://example.test/release",

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

Executable file

)         self.assertNotIn("--json", skillhub_command)         self.assertIn("https://api.skillhub.cn", skillhub_command)

self.assertNotIn("--json", skillhub_command)         self.assertIn("https://api.skillhub.cn", skillhub_command)         self.assertIn(

self.assertIn("https://api.skillhub.cn", skillhub_command)

self.config.clawhub,                 owner="configured-owner",                 endpoint="https://clawhub.example",

owner="configured-owner",                 endpoint="https://clawhub.example",                 package="clawhub@1.2.3",

endpoint="https://clawhub.example",

package="clawhub@1.2.3",             ),             skillhub=replace(self.config.skillhub, endpoint="https://skillhub.example"),

),             skillhub=replace(self.config.skillhub, endpoint="https://skillhub.example"),         )

skillhub=replace(self.config.skillhub, endpoint="https://skillhub.example"),

[Credential-related source omitted]

for command in clawhub_commands:             self.assertEqual(                 command[command.index("--registry") + 1], "https://clawhub.example"

self.assertEqual(                 command[command.index("--registry") + 1], "https://clawhub.example"             )

command[command.index("--registry") + 1], "https://clawhub.example"

if command[0] == "/bin/skillhub":                 self.assertEqual(                     command[command.index("--host") + 1], "https://skillhub.example"

self.assertEqual(                     command[command.index("--host") + 1], "https://skillhub.example"                 )

command[command.index("--host") + 1], "https://skillhub.example"

[Credential-related source omitted]

[Credential-related source omitted]

"workbuddy",             "verified",             url="https://workbuddy.example/releases/1.1.0",

"verified",             url="https://workbuddy.example/releases/1.1.0",         )

url="https://workbuddy.example/releases/1.1.0",

"doubao",             "submitted",             url="https://doubao.example/submissions/123",

"submitted",             url="https://doubao.example/submissions/123",         )

url="https://doubao.example/submissions/123",

Executable file

[Credential-related source omitted]

"number": 77,                     "title": "feat: Improve course authoring and continuation",                     "html_url": "https://github.com/example/skills/pull/77",

"title": "feat: Improve course authoring and continuation",                     "html_url": "https://github.com/example/skills/pull/77",                     "user": {"login": "author"},

"html_url": "https://github.com/example/skills/pull/77",

"number": 77,                     "title": "feat: Improve course authoring and continuation",                     "html_url": "https://github.com/example/skills/pull/77",

"title": "feat: Improve course authoring and continuation",                     "html_url": "https://github.com/example/skills/pull/77",                     "user": {"login": "author"},

"html_url": "https://github.com/example/skills/pull/77",

Executable file

"number": number,             "title": title,             "html_url": f"https://github.com/example/skills/pull/{number}",

"title": title,             "html_url": f"https://github.com/example/skills/pull/{number}",             "user": {"login": "author"},

"html_url": f"https://github.com/example/skills/pull/{number}",

files=None,     ):         def fetch(endpoint: str, *, paginate: bool = False):

):         def fetch(endpoint: str, *, paginate: bool = False):             if endpoint == "repos/example/skills":

def fetch(endpoint: str, *, paginate: bool = False):

for number in (100, 200, 300):             self.assertEqual(text.count(f"[#{number}]"), 1)             self.assertIn(f"https://github.com/example/skills/pull/{number}", text)

self.assertEqual(text.count(f"[#{number}]"), 1)             self.assertIn(f"https://github.com/example/skills/pull/{number}", text)         self.assertIn(" by @author", text)

self.assertIn(f"https://github.com/example/skills/pull/{number}", text)

self.assertIn("### Additional commits", text)         self.assertIn("### Unmerged preview changes", text)         self.assertIn(f"https://github.com/example/skills/commit/{direct}", text)

self.assertIn("### Unmerged preview changes", text)         self.assertIn(f"https://github.com/example/skills/commit/{direct}", text)         self.assertEqual(snapshot["pr_count"], 0)

self.assertIn(f"https://github.com/example/skills/commit/{direct}", text)

with self.assertRaisesRegex(ValueError, "differs from the source"):             release_notes.collect_changes(                 "https://github.com/example/skills.git",

release_notes.collect_changes(                 "https://github.com/example/skills.git",                 head,

"https://github.com/example/skills.git",

"author": "sunner",                     "author_is_login": True,                     "url": "https://github.com/ai-shifu/skills/pull/196",

"author_is_login": True,                     "url": "https://github.com/ai-shifu/skills/pull/196",                     "merged_at": "2026-10-20T08:25:29Z",

"url": "https://github.com/ai-shifu/skills/pull/196",

"author": "sunner",                     "author_is_login": True,                     "url": "https://github.com/ai-shifu/skills/pull/195",

"author_is_login": True,                     "url": "https://github.com/ai-shifu/skills/pull/195",                     "merged_at": "2026-10-20T08:21:48Z",

"url": "https://github.com/ai-shifu/skills/pull/195",

"### Bug Fixes\n\n"             "- reject invalid course imports before changing content "             "([#195](https://github.com/ai-shifu/skills/pull/195)) by @sunner\n\n"

"- reject invalid course imports before changing content "             "([#195](https://github.com/ai-shifu/skills/pull/195)) by @sunner\n\n"             "### Maintenance\n\n"

"([#195](https://github.com/ai-shifu/skills/pull/195)) by @sunner\n\n"

"### Maintenance\n\n"             "- flow version to v2.1.1 "             "([#196](https://github.com/ai-shifu/skills/pull/196)) by @sunner\n\n"

"- flow version to v2.1.1 "             "([#196](https://github.com/ai-shifu/skills/pull/196)) by @sunner\n\n"             "[Full comparison](https://github.com/ai-shifu/skills/compare/"

"([#196](https://github.com/ai-shifu/skills/pull/196)) by @sunner\n\n"

"[Full comparison](https://github.com/ai-shifu/skills/compare/"

Executable file

return eval(  # noqa: S307 - AST and execution context are explicitly restricted.

[Credential-related source omitted]

[Credential-related source omitted]

'mkdir -p release-notes\n'                     'printf "%s\\n" "Release notes" > release-notes/release-notes.md\n'                     'printf "%s\\n" "https://github.com/example/skills/releases/tag/v1.2.3"\n'

'printf "%s\\n" "Release notes" > release-notes/release-notes.md\n'                     'printf "%s\\n" "https://github.com/example/skills/releases/tag/v1.2.3"\n'                     'exit "$PUBLICATION_EXIT_CODE"\n'

'printf "%s\\n" "https://github.com/example/skills/releases/tag/v1.2.3"\n'

[Credential-related source omitted]

[Credential-related source omitted]

[Credential-related source omitted]

Executable file

[Credential-related source omitted]

1. Read `https://api.github.com/repos/ai-shifu/skills/releases/latest`, or the Release for the requested tag. Record the tag, Release URL, asset name, size, and available SHA-256 digest. Confirm that the Release is published and stable. 2. 

1. Open `https://open.workbuddy.cn/expert/all` in the authenticated browser, under `发布管理` → `专家`. If login is needed, let the user complete it; never store credentials in this skill or its outputs.