Token导航 LogoToken导航TokenDH.com
待分类external-servicegithub未标认证来源可访问许可证需确认审计通过

journal-reflection日记反思

Agent Skill

journal-reflection 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

514

周安装

21

GitHub Stars

公开资料未说明

下载量

166
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:journal-reflection(日记反思)
来源仓库:https://github.com/gasserane/personal-skills
仓库路径:skills/journal-reflection
安装命令:
npx skills add https://github.com/gasserane/personal-skills --skill journal-reflection
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/gasserane/personal-skills --skill journal-reflection

简介

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 中围绕仓库状态、代码变更或协作事项进行整理。
  • 通过 npx skills add 命令从指定仓库安装并使用。
  • 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。
  • journal-reflection 属于待分类类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Journal Reflection

Structured reflection without the usual journaling bloat. Four modes. Pick one and run it tight.

When to use

Trigger when the user asks to reflect, review, or capture learning. Not for morning briefings — route to daily-brief for that.

Mode selection

Ask once, in one prompt, which mode applies. Default to end-of-day if the user says "journal" without specifying.

  1. End-of-day (5 min): today in 3 questions
  2. End-of-week (15 min): week in 5 questions plus a pattern scan
  3. Decision review: one past decision in 4 questions
  4. Learning log: one new framework or concept, linked to what Ane already knows

Prompts by mode

End-of-day

Ask these three, one at a time, waiting for each answer:

  1. What got done today that mattered? (not a task list — what *moved*)
  2. What got in the way, and what did it cost?
  3. What does tomorrow need from you that today did not get?

Close with: one commitment for tomorrow, one sentence, verb-first.

End-of-week

Ask these five, one at a time:

  1. Against this week's stated goals, what is the honest progress? (percentage-free language)
  2. What patterns showed up this week that also showed up last week?
  3. What did you learn about the work, or about yourself doing the work?
  4. What will you change next week? Specific, small, testable.
  5. What deserves to be named as a win that you have not named yet?

Close with: one pattern-to-watch next week.

Decision review

Ask:

  1. What was the decision, and when did you make it?
  2. What did you know at the time, and what were you guessing?
  3. What happened in the days or weeks since?
  4. What would you do differently if the decision arrived today?

Close with: one rule to carry forward, written as an if-then.

Learning log

Ask:

  1. What is the concept or framework? Cite author and year if it has one.
  2. What problem does it solve that older frameworks do not?
  3. How does it connect to something Ane already uses?
  4. Where will Ane apply it first?

Output format

Write Obsidian-compatible markdown with YAML frontmatter:

---
date: YYYY-MM-DD
type: [daily | weekly | decision-review | learning-log]
tags: [journal, ...domain tags]
dg-publish: false
---

Use [[wiki-links]] for people, concepts, frameworks, and projects. Default dg-publish to false. If the user says the note is publishable, set to true and ask for the permalink stub.

Body structure: one H2 per prompt, the answer as plain prose under it, no bullets unless the content is a list.

File placement

If filesystem MCP to MELdigitalgarden is connected:

  • End-of-day: src/site/notes/Journal/Daily/YYYY-MM-DD.md
  • End-of-week: src/site/notes/Journal/Weekly/YYYY-Www.md
  • Decision review: src/site/notes/Journal/Decisions/YYYY-MM-DD-<slug>.md
  • Learning log: src/site/notes/Journal/Learning/<slug>.md

If MCP is not connected, return the markdown as an artifact and name the target path so the user can save it manually.

Writing rules

Follow CLAUDE.md house style. One exception: in journal answers, passive and reflective phrasing is acceptable when Ane's answer is itself reflective. Do not rewrite her words into active voice — preserve her voice exactly.

Limitations

Do not invent reflections. Do not auto-answer the prompts. Prompts go to the user and wait. If the user skips a prompt, mark the section (skipped) and move on. Do not lecture.

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

32.84%
按下载量换算55

Claude

27.86%
按下载量换算46

Cursor

20.52%
按下载量换算34

Gemini CLI

8.89%
按下载量换算15

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

external-service

该 Skill 可能调用第三方服务、云服务或外部模型 API,使用前需要确认账号、额度、数据发送范围和服务条款。

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