Token导航 LogoToken导航TokenDH.com
前端设计只读github未标认证来源可访问许可证需确认审计异常

note-creator笔记创建者

Agent Skill

note-creator 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

212

周安装

9

GitHub Stars

8

下载量

74
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:note-creator(笔记创建者)
来源仓库:https://github.com/lingengyuan/my-skills
仓库路径:skills/note-creator
安装命令:
npx skills add https://github.com/lingengyuan/my-skills --skill note-creator
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/lingengyuan/my-skills --skill note-creator

简介

用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息。

  • 适合围绕仓库状态、代码变更或协作事项进行整理,提升开发效率。
  • 可结合来源仓库和原始 README 核验具体用法,确保适用性。
  • 安装方式:通过 npx 从 GitHub 仓库添加,支持 Codex、Claude、Cursor、Gemini CLI。
  • 注意:安装前建议确认权限范围和维护状态,避免触发不必要的联网或文件操作。

SKILL.md

note-creator (Orchestrator Skill)

Purpose

Given a short user prompt (and optional raw context), this skill MUST:

  1. Classify intent and decide a minimal artifact plan (default: md only)
  2. Select a destination folder and stable naming
  3. Invoke underlying format skills with runtime context
  4. WRITE all generated artifacts to disk following a strict contract

This skill is an ORCHESTRATOR. Generating content without writing files is considered a FAILURE.


Dependencies

This skill MUST delegate format correctness to:

  • obsidian-markdown/SKILL.md
  • json-canvas/SKILL.md
  • obsidian-bases/SKILL.md

Inputs

  • user_prompt: a short sentence or paragraph describing what to record
  • optional_context_files (optional): raw markdown or text provided by the user

Output Contract

CRITICAL: Output paths are relative to CURRENT WORKING DIRECTORY (CWD)

  • The CWD when the skill is invoked is the base (e.g., F:/Project/Obsidian/)
  • All outputs MUST be written to: <cwd>/outputs/<folder>/<title>/
  • DO NOT write to the skill's own directory
  • When computing paths, always use the current working directory as the base

All outputs MUST be written to:

outputs///

  • note.md (required)
  • diagram.canvas (optional)
  • table.base (optional)
  • meta.json (required)

Rules:

  • MUST come from rules/folders.md
  • - Writing files is mandatory

Algorithm (Strict Execution Checklist)

1) Read Inputs

  • Read user_prompt
  • Read any optional_context_files

2) Classify Intent

Produce a STRICT classification JSON according to rules/classify.intent.md. The JSON MUST include:

  • title
  • folder
  • diagram_type
  • artifact_plan (must include "md")
  • tags (3–8)
  • properties (category, created, modified, source)

If the request is a COMPARISON and "base" is included, the JSON MUST ALSO include:

  • base_mode = "comparison"
  • comparison_items: an array of items to compare (slug + display_name + optional fields)

3) Compute Output Paths

  • CRITICAL: Get the current working directory (CWD) first
  • CWD is the directory where the user invoked the skill (e.g., F:/Project/Obsidian/)
  • Compute all paths as relative to CWD:

- root_dir = /outputs/// - note_path = /outputs///note.md - canvas_path = /outputs///diagram.canvas - base_path = /outputs///table.base - meta_path = /outputs///meta.json - compare_dir = /outputs///compare/ (comparison only)

  • When writing files, use these full paths or ensure you're in the CWD

4) Generate and WRITE note.md (MANDATORY)

  • Invoke obsidian-markdown skill using templates/note.md.prompt
  • The invocation MUST include runtime context:

- title - folder - tags - properties - root_dir (/outputs///)

  • The generated markdown MUST be written to disk at: /outputs///note.md

5) Generate and WRITE diagram.canvas (CONDITIONAL)

  • If "canvas" is present in artifact_plan:

- Select ONE canvas layout template based on diagram_type and content intent: - If diagram_type == "sequence": - If the note intent is documentation, explanation, sharing, usage instructions, or public-facing content: -> use templates/canvas.sequence.compact.md - Otherwise (internal notes, debugging, design reasoning): -> use templates/canvas.sequence.detailed.md - If diagram_type == "flowchart": -> use templates/canvas.flowchart.md - If diagram_type == "artifact": -> use templates/canvas.artifact.md - If diagram_type == "architecture": -> use templates/canvas.architecture.md - If diagram_type == "none": -> DO NOT generate diagram.canvas (remove "canvas" from artifact_plan) - Invoke json-canvas skill using templates/canvas.prompt - The invocation MUST include runtime context: - title - folder - diagram_type - artifact_plan - root_dir (/outputs///) - canvas_template (the FULL CONTENT of the selected templates/canvas.*.md) - The generated canvas MUST be written to disk at: /outputs///diagram.canvas

6) Generate and WRITE table.base (CONDITIONAL)

  • If "base" is present in artifact_plan: 6.1 Comparison Base Mode (PREFERRED for comparisons) 6.2 Generic Base Mode (fallback)

- If base_mode == "comparison" AND comparison_items is present and non-empty: - IMPORTANT: - If base_mode == "comparison" BUT comparison_items is missing or empty, this is a classification error. - In this case, DO NOT fall back to generic base generation. - The execution MUST either: - FAIL explicitly; OR - Remove "base" from artifact_plan and continue with md-only output. 1. CREATE compare_dir: /outputs///compare/ 2. For EACH item in comparison_items: - Generate ONE markdown file: /outputs///compare/.md - Use templates/compare.item.md as a strict template. - Fill frontmatter fields so Base can render columns: - item_type: "skill" - name - 输入 - 输出 - 定位 - 产物 - 边界 - tags (include "对比") - source/created/modified (from properties) - IMPORTANT: - These compare/*.md files are the "rows" of the Base. - Do NOT put compare rows into note.md only; they MUST exist as files. 3. Invoke obsidian-bases skill using templates/base.comparison.prompt The invocation MUST include runtime context: - title - folder - root_dir (/outputs///) - compare_dir (/outputs///compare/) - properties (category/created/modified/source) 4. The generated base MUST be written to disk at: /outputs///table.base - Else IF base_mode is NOT "comparison": - Invoke obsidian-bases skill using templates/base.prompt - The invocation MUST include runtime context: - title - folder - artifact_plan - root_dir (/outputs///) - The generated base MUST be written to disk at: /outputs///table.base

7) Generate and WRITE meta.json (MANDATORY)

  • Generate meta.json according to rules/output.contract.md
  • Include comparison metadata when applicable:

- base_mode - compare_dir - comparison_items

  • Write meta.json to: /outputs///meta.json

8) Execution Summary

  • Output a short summary listing generated files and their final paths

File Writing Rules (CRITICAL)

  • Every generated artifact MUST be written to disk.
  • Outputting content only in the conversation is NOT allowed.
  • Missing file writes invalidate the execution.
  • Always write files relative to the current working directory, not the skill directory

Hard Constraints

General

  • artifact_plan MUST include "md"
  • folder MUST be from rules/folders.md
  • tags count MUST be 3–8 and include at least one domain tag

note.md

Must include:

  • YAML frontmatter (tags, category, created, modified)
  • Summary section
  • ≥3 structured sections
  • Examples (code blocks when applicable)
  • ≥5 Caveats / Notes

diagram.canvas (if generated)

  • Text nodes MUST use "type":"text" and the "text" field
  • MUST NOT use "label" for text nodes
  • ≥6 nodes, ≥5 edges
  • Must include: user request, intent classification, and persistence to root_dir

table.base (if generated)

  • For comparison mode:

- MUST scope sources to compare_dir ONLY - MUST include only markdown files - MUST prefer frontmatter columns from compare/*.md

  • For generic mode:

- MUST scope sources to root_dir only - MUST exclude non-markdown files


Examples

See examples/ directory:

  • e2e-docker.md
  • e2e-arch.md
  • e2e-table.md

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

38.83%
按下载量换算29

Claude

28.59%
按下载量换算21

Cursor

19.22%
按下载量换算14

Gemini CLI

9.34%
按下载量换算7

安全审计

Gen Agent Trust Hub

未通过

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。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。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