Token导航 LogoToken导航TokenDH.com
研究检索只读github未标认证来源可访问许可证需确认审计通过

ankianki 开发

Agent Skill

用于辅助文档、README、Markdown、说明文和内容稿件的整理与改写。它适合让 Agent 提炼结构、补齐章节、统一术语、检查链接或把零散材料整理成可读文档。使用时应保留项目已有事实、命令和路径,不要把未确认的信息写成确定结论;涉及对外文案时,还需要控制语气,避免过度营销或夸大能力。

总安装

2,136

周安装

89

GitHub Stars

20

下载量

712
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:anki(anki 开发)
来源仓库:https://github.com/terkelg/anki-markdown
仓库路径:skills/anki
安装命令:
npx skills add https://github.com/terkelg/anki-markdown --skill anki
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/terkelg/anki-markdown --skill anki

简介

用于辅助文档、Markdown 和内容稿件的整理与改写。

  • 可提炼结构、补齐章节或检查链接,提升文档质量。
  • 保留项目已有事实和命令,不将未确认信息写成确定结论。
  • 对外文案需注意语气,避免过度营销或夸大能力。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。
  • anki 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Anki Flashcard Management

Use the anki.sh script to create and manage flashcards via AnkiConnect.

ANKI="$(dirname "$SKILL_PATH")/scripts/anki.sh"

Preflight:

  • Ensure Anki desktop is running with AnkiConnect enabled
  • Ensure curl and jq are available in PATH

CRITICAL: User Approval Required

NEVER add cards to Anki without explicit user approval.

Before adding any cards:

  1. Display all proposed cards in full (front + back)
  2. Ask user to confirm: "Ready to add these cards?"
  3. Only proceed after explicit "yes" / approval

This is non-negotiable. Users must review card quality before adding.

For potentially destructive or irreversible actions:

  • delete: ask explicit confirmation before deleting notes
  • update: confirm exact fields/IDs before updating notes
  • rate: show card + suggested rating, then wait for user confirmation before submitting

Guidelines

  • MUST use the Anki Markdown note type for basic cards or Anki Markdown Cloze for cloze deletions
  • Use markdown for both front and back fields
  • Test exactly one fact or idea per card
  • Make questions atomic and simple
  • Keep questions and answers short, clear, and unambiguous
  • Ensure each card has one unique, specific answer
  • Avoid long lists or multiple items in one card
  • Add context or hints so your future self understands
  • Personalize with examples or mnemonics if helpful
  • Allow multiple simple cards for important concepts
  • Create multiple variants of important questions from different angles
  • Connect cards to personal goals and projects to avoid "orphan" facts
  • Always use syntax highlighting for code (see Code Snippets section)
  • When creating multiple related cards, use bulk operations for efficiency

Choosing Card Type

Use Anki Markdown (Front/Back) for question-answer pairs where you write a specific question. Use Anki Markdown Cloze (Text/Extra) for fill-in-the-blank cards where hiding parts of a statement is more natural than writing a question.

Cloze Cards

Use Anki Markdown Cloze for fill-in-the-blank cards. Fields are Text and Extra.

Syntax

Three forms:

  • {{c1::answer}} - shows [...] on front
  • {{c1::answer::hint}} - shows [hint] on front
  • {{c1::answer::blur}} - shows blurred content on front (reveals shape but not text)

Full markdown works inside cloze deletions: bold, italic, highlights, inline code with language tags, links, images, code blocks, and tables.

Multiple Cards

Each cloze number (c1, c2, c3...) generates a separate card. Use different numbers to test different facts from the same note:

{"Text": "{{c1::HTML}} provides structure, {{c2::CSS}} provides style."}

This produces 2 cards. Card 1 hides "HTML", card 2 hides "CSS".

Using the same number multiple times hides all instances on the same card:

{"Text": "{{c1::React}} and {{c1::Vue}} are frontend frameworks."}

This produces 1 card that hides both "React" and "Vue".

Nested Cloze

Cloze deletions can be nested. The outer cloze hides everything including the inner:

{"Text": "{{c1::Canberra was {{c2::founded}}}} in 1913."}

Card 1 hides "Canberra was founded". Card 2 only hides "founded".

Inline Code in Cloze

Inline code with language tags works inside cloze. Note the triple closing braces (}}}) when code uses {lang} syntax:

{"Text": "Use {{c1::`querySelector()`{js}}} to find DOM elements."}

Extra Field

The Extra field appears on the back of every card. Use it for context, examples, or references. It supports full markdown including code blocks and callouts.

Cloze Examples

# Basic cloze
$ANKI add "MyDeck" "Anki Markdown Cloze" '{"Text":"The {{c1::CPU}} executes {{c2::instructions}}.","Extra":"Basic computer architecture."}' --tags "cs"

# Cloze with hint
$ANKI add "MyDeck" "Anki Markdown Cloze" '{"Text":"{{c1::JavaScript::language}} was created in {{c2::1995::year}}.","Extra":"Created by Brendan Eich."}' --tags "history"

# Cloze with blur
$ANKI add "MyDeck" "Anki Markdown Cloze" '{"Text":"The speed of light is {{c1::299,792,458 m/s::blur}}.","Extra":"Often approximated as 3 x 10^8 m/s."}' --tags "physics"

Front Field Format

Use markdown on the front to highlight key terms the question is about:

  • Use bold or ==highlight== for the main concept being tested
  • Use inline code for technical terms, functions, or syntax

Example: What does **vertex_index** return in WGSL?

Back Field Format

Use this format for the Back field:

  1. Bold one-liner with the direct answer
  2. Blank line
  3. Additional context in normal text (optional)

Formatting for Fast Skimming

  • Use ==highlighted== to make key terms stand out at a glance
  • Use bold for critical concepts within explanations
  • Use syntax highlighting for all code blocks including inline
  • Optimize for quick visual scanning—readers should grasp the core idea instantly

Keyboard Shortcuts

Use <kbd> tags for keyboard keys:

Press <kbd>Cmd</kbd>+<kbd>Shift</kbd>+<kbd>P</kbd> to open the command palette.

Code Snippets

Always use syntax highlighting:

  • Fenced blocks: Use triple backticks with language (js,python, ```rust, etc.)
  • Inline code: Use ` code{lang} syntax (e.g., const x = 1{js} `)

Advanced features:

  • Line highlighting: ```js {2} to highlight line 2
  • Word highlighting: ```js /pattern/ to highlight specific words
  • Focus mode: Add // [!code focus] comment to focus specific lines

Callouts

Use GitHub-style alerts sparingly for important notes:

> [!TIP]
> Brief helpful insight here.

> [!NOTE]
> Useful context the learner should know.

> [!WARNING]
> Common mistake or gotcha to avoid.

Example

Front: What does `passEncoder.draw(3)`{js} do with **no vertex buffer**?
Back:

**Invokes the ==vertex shader== 3 times.**

The shader uses `@builtin(vertex_index)`{wgsl} to generate positions ==procedurally==.

> [!TIP]
> Useful for fullscreen quads without geometry data.

Common Operations

Run via $ANKI <action> [args...]:

ActionDescriptionExample
syncTrigger AnkiWeb sync$ANKI sync
decksList decks$ANKI decks --stats
modelsList note types$ANKI models
fieldsFields for a model$ANKI fields "Anki Markdown"
findSearch notes$ANKI find "deck:Spanish tag:verb"
infoNote details$ANKI info 1234 5678
addAdd one note$ANKI add "My Deck" "Anki Markdown" '{"Front":"Q","Back":"A"}' --tags "t1 t2"
add-bulkAdd many notes$ANKI add-bulk "My Deck" "Anki Markdown" '[{"Front":"Q1","Back":"A1"}]'
updateUpdate fields$ANKI update 1234 '{"Front":"New Q"}'
deleteDelete notes$ANKI delete 1234 5678
dueGet due cards$ANKI due "My Deck" --limit 5
reviewShow card$ANKI review 1234
rateRate card (1-4)$ANKI rate 1234 3
tagsList tags$ANKI tags --pattern "verb"

Typical workflow

$ANKI sync                          # sync first
$ANKI decks                         # check available decks
$ANKI models                        # verify required model exists
$ANKI fields "Anki Markdown"        # verify field names for selected model
$ANKI find "deck:MyDeck front:test" # check for duplicates
$ANKI add "MyDeck" "Anki Markdown" '{"Front":"...","Back":"..."}'
$ANKI sync                          # sync after changes

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

37.19%
按下载量换算265

Claude

31.92%
按下载量换算227

Cursor

19.04%
按下载量换算136

Gemini CLI

8.47%
按下载量换算60

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills