Agent skill

Auto Skill

by Toolsai in Toolsai/auto-skill

在使用者明確同意的前提下,管理 Agent 任務的本地知識庫與跨技能經驗索引;當需要讀取、整理、驗證或回寫可重用的工作經驗時使用。絕不自動修改 IDE 或 Agent 全域規則。

MITAuto-check passedKnowledge Management

Install Auto Skill

skills CLI
$ npx skills add Toolsai/auto-skill --skill auto-skill -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install Toolsai/auto-skill auto-skill --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
auto-skill
GitHub stars
193
Token cost
~991 tokens
SKILL.md length
208 words
Files
14 (incl. assets)
Skills in repo
1
Repo updated
First seen
Licence
MIT

At a glance

在使用者明確同意的前提下,管理 Agent 任務的本地知識庫與跨技能經驗索引;當需要讀取、整理、驗證或回寫可重用的工作經驗時使用。絕不自動修改 IDE 或 Agent 全域規則。

  • Works in 7 steps: 安全初始化(只讀) → 對話內快取(不對用戶展示) → 每回合先抽取關鍵詞(不讀檔) → …
  • Knowledge Management work in your project
  • SKILL.md covers 核心循環(Step 1–5), 記錄判斷準則, 條目格式 and 存儲路徑, plus 2 more sections
  • Calls npm

What it does

Auto Skill is an agent skill from Toolsai/auto-skill. 在使用者明確同意的前提下,管理 Agent 任務的本地知識庫與跨技能經驗索引;當需要讀取、整理、驗證或回寫可重用的工作經驗時使用。絕不自動修改 IDE 或 Agent 全域規則。

Its SKILL.md is about 990 tokens, which your agent loads only when the skill is triggered. The skill folder holds 18 other files, including assets (for example `.github/workflows/validate.yml`, `README.md` and `experience/_index.json`).

It sits in Knowledge Management. The licence is MIT.

When your agent uses it

  • Knowledge Management work in your project

Example prompts

  • “/auto-skill”

Requirements

  • Node.js

Workflow steps

7 steps, taken from the step headings in SKILL.md.

  1. 安全初始化(只讀)
  2. 對話內快取(不對用戶展示)
  3. 每回合先抽取關鍵詞(不讀檔)
  4. 判斷是否話題切換(不讀檔)
  5. 跨技能經驗讀取(強制規則,不受話題切換影響)
  6. 只在話題切換時讀取知識庫(knowledge-base)
  7. 任務結束:主動記錄(最重要!)

What it can do on your machine

Read from SKILL.md and the folder at commit f4e042c. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • npm

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use npm, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Auto Skill loads about 991 tokens when it runs. Until then it costs about 25 tokens; SKILL.md has 208 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~25
When it runs · the whole SKILL.md, loaded when a task matches
~991

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from Toolsai/auto-skill at commit f4e042c, republished under its MIT licence (© Toolsai). 208 words, ~991 tokens.

Download SKILL.mdSave it as .claude/skills/auto-skill/SKILL.md (or your agent's skills folder). This skill also uses 13 other files; get the full folder from GitHub.
name
auto-skill
description
在使用者明確同意的前提下,管理 Agent 任務的本地知識庫與跨技能經驗索引;當需要讀取、整理、驗證或回寫可重用的工作經驗時使用。絕不自動修改 IDE 或 Agent 全域規則。
license
MIT

Auto-Skill 自進化知識系統

核心循環(Step 1–5)

你必須在每一輪對話中遵循以下核心循環:

0. 安全初始化(只讀)

首次觸發 auto-skill 時:

  1. 確認知識庫的根目錄與寫入範圍;若未設定或不清楚,先詢問使用者。
  2. 只讀檢查索引與內容檔案是否存在、格式是否有效。
  3. 絕不修改 IDE/Agent 全域規則檔,也不建立全域指令或啟動協議。
  4. 不要宣稱已完成任何寫入;只有在工具回報成功且重新讀取驗證後才能回報。
0. 對話內快取(不對用戶展示)

在同一對話串中維護以下快取:

  • last_keywords
  • last_topic_fingerprint
  • last_index_lastUpdated
  • last_matched_categories
  • last_used_skills(本回合用到的非 auto-skill 技能清單)
  • missing_experience_skills(experience 未命中的技能)
  • loaded_experience_skills(本對話已讀取過經驗的 skill-id)
寫入安全邊界
  • 只有在使用者明確同意後才寫入知識庫或經驗庫。
  • 寫入前移除 API keys、tokens、密碼、cookies、私鑰與不必要的原始對話內容。
  • 只寫入使用者核准的資料根目錄;拒絕絕對路徑、.. 路徑與指向根目錄外的符號連結。
  • 不要把個人記憶寫回已安裝的 Skill 目錄;使用者資料與 Skill 套件必須分離。
  • 更新 JSON 索引前先驗證格式,並使用暫存檔加原子替換,避免部分寫入破壞索引。
1. 每回合先抽取關鍵詞(不讀檔)
  • 從當前用戶訊息抽取 3–8 個核心名詞/短語(去重、統一大小寫)。
  • 生成 topic_fingerprint = 前 3 個關鍵詞。
2. 判斷是否話題切換(不讀檔)

當出現以下任一條件,視為話題切換:

  • 明確轉折詞:例如「另外」「改成」「換成」「再來」「順便」
  • 本回合關鍵詞與 last_keywords 差異 >= 40%
  • 用戶明確要求新增/修改分類
3. 跨技能經驗讀取(強制規則,不受話題切換影響)

只要本回合使用了任何「非 auto-skill」技能:

  • 若該 skill-id 已存在於 loaded_experience_skills,本回合不重讀、不重複提示
  • 否則必須執行以下步驟:
    1. 讀取 experience/_index.json
    2. 若找到對應 skill-id,必須載入該經驗檔 experience/skill-[skill-id].md
    3. 將該 skill-id 加入 loaded_experience_skills
    4. 回覆中必須提示:我已讀取經驗:skill-xxx.md
    5. 若 experience/_index.json 沒有該技能,記錄到 missing_experience_skills
4. 只在話題切換時讀取知識庫(knowledge-base)

若是本對話第一次回合或判定話題切換,才執行以下步驟:

  • 讀取 knowledge-base/_index.json
  • 以本回合關鍵詞匹配所有分類 keywords
  • 匹配到多少分類就讀多少分類(不做優先級排序)
  • 若沒有匹配分類,依「動態分類」流程處理
  • 若本回合有讀取任何分類檔,回覆中需加入一行提示: 我已讀取知識庫:design-layout.md, frontend-dev.md (以實際讀取檔名替換,逗號分隔)

若不是話題切換,沿用 last_matched_categories,不重讀索引與分類檔。

5. 任務結束:主動記錄(最重要!)

任務明顯已完成:你判斷本回合已高完成且值得記錄時 觸發詞:用戶表達對任務滿意時

你必須執行以下步驟:

  1. 總結經驗:用一句話提煉本次解決方案的精華
  2. 判斷價值:這個經驗下次能幫用戶省時間嗎?
  3. 主動詢問:必須說出類似這樣的話:

    「這次我們解決了 [問題描述],我想把這個經驗記錄到你的知識庫,下次遇到類似問題時可以直接參考。你覺得可以嗎?」

  4. 執行記錄:用戶同意後,依下列規則寫入並更新索引:
    • 跨技能經驗:若本回合使用非 auto-skill,且該技能在 experience 中不存在或有新技巧 → 寫入 experience/skill-[skill-id].md,更新 experience/_index.json
    • 一般知識:若為通用流程/偏好/解法 → 寫入 knowledge-base/[category].md,更新 knowledge-base/_index.json

強制規則:缺少經驗時必問 若本回合使用了非 auto-skill 技能,且該技能不在 experience/_index.json:

  • 任務結束時必須主動詢問是否記錄本次使用經驗
  • 詢問語句需明確指向該技能,例如:

    「這次使用了 remotion-best-practices,但經驗庫沒有紀錄。我可以把這次的做法記錄下來嗎?」


記錄判斷準則

核心問題:這東西下次能讓用戶省時間嗎?

General(knowledge-base)

應該記錄(general):

  • ✅ 可重用的流程與決策步驟(跨領域通用的操作順序/判斷流程)
  • ✅ 高成本的錯誤與修正路徑(犯錯會浪費大量時間的情況)
  • ✅ 關鍵參數/設定/前置條件(一變就影響結果的要素)
  • ✅ 使用者偏好與風格規則(語氣、格式、設計風格、輸出結構)
  • ✅ 多次嘗試才成功的方案(包含失敗原因與成功條件)
  • ✅ 可套用的模板/清單/格式(會反覆使用的輸出樣式)
  • ✅ 外部依賴或資源位置(檔案路徑、工具、素材)

不應記錄(general):

  • ❌ 一問一答、沒有可重用流程
  • ❌ 純概念解釋(沒有具體做法或判斷標準)
  • ❌ 沒有具體上下文、不可復用的結論
Experience(非 auto-skill 經驗)

應該記錄(experience):

  • ✅ 使用該技能時踩到的坑與解法(含錯誤訊息/定位方式)
  • ✅ 影響結果的關鍵參數或配置(如 spring 參數、fps、duration)
  • ✅ 可重用的模板/提示詞/工作流程(可直接套用)
  • ✅ 依賴或資產路徑(字體、圖片、專案入口、模組位置)
  • ✅ 需要特定順序或技巧才成功的步驟(例如先初始化再覆蓋)

不應記錄(experience):

  • ❌ 純理論或概念性解釋(留在 knowledge-base)
  • ❌ 沒有可重現步驟的結論
  • ❌ 一次性、不可重用的操作

條目格式

knowledge-base 條目格式
markdown
## 🔧 [簡短標題]
**日期:** YYYY-MM-DD
**情境:** 一句話描述使用場景
**最佳實踐:**
- [重點 1]
- [重點 2] - 參數說明和調整指南
experience 條目格式
markdown
## 🔧 [問題/技巧標題]
**日期:** YYYY-MM-DD
**技能:** [skill-id]
**情境:** 一句話描述本次問題
**解法:**
- 具體步驟 1
- 具體步驟 2
**關鍵檔案/路徑:**
- /path/to/file
**keywords:** keyword1, keyword2, keyword3

存儲路徑

  • 知識索引:knowledge-base/_index.json
  • 知識內容:knowledge-base/[category].md
  • 經驗索引:experience/_index.json
  • 經驗內容:experience/skill-[skill-id].md

動態分類(僅 knowledge-base)

當用戶的問題不屬於現有分類時:

  1. 建議創建新分類
  2. 詢問用戶分類名稱和關鍵詞
  3. 創建新的 .md 文件並更新 _index.json

QMD 升級(未來)

當知識庫條目 > 50 條時,主動建議用戶安裝 QMD:

bash
npm install -g qmd && qmd collection add knowledge-base --name auto-skill && qmd embed

安裝後,改用 qmd_query 工具進行語義檢索。

© Toolsai, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 13 other files (assets) in the repository root of Toolsai/auto-skill.

  • SKILL.md
  • .github/workflows/validate.yml
  • LICENSE
  • README.md
  • assets/auto-skill-flow.webp
  • assets/auto-upload-knowlege.webp
  • experience/_index.json
  • experience/skill-remotion-best-practices.md
  • knowledge-base/_index.json
  • knowledge-base/backend-dev.md
  • knowledge-base/design-layout.md
  • knowledge-base/frontend-dev.md
  • knowledge-base/workflow.md
  • knowledge-base/writing.md

Open the folder on GitHubat commit f4e042c

Compare with similar skills

Auto Skill next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Auto Skill compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Auto Skill this skillToolsai/auto-skill193—~991Automated safety check: PassMIT
Baoyu URL To Markdownsdyckjq-lab/llm-wiki-skill2.5k3 repos~3.2kAutomated safety check: PassNone
Logseq Review Workflow Evallogseq/logseq45k—~1kAutomated safety check: PassAGPL-3.0
Obsidian CLIAtmosphere/atmosphere3.8k13 repos~795Automated safety check: PassApache-2.0
Esm Cjs Risk Scanlogseq/logseq45k—~3.3kAutomated safety check: PassAGPL-3.0
Capture Conversationoutline/outline41k—~474Automated safety check: PassCustom licence

Similar skills

  • Baoyu URL To Markdown

    sdyckjq-lab/llm-wiki-skill

    Fetch any URL and convert to markdown using Chrome CDP. An agent skill from sdyckjq-lab/llm-wiki-skill.

    2.5k GitHub starsUsed in 3 repos~3.2k tokens
    Knowledge ManagementAuto-check passed
  • Compare two revisions of the Logseq logseq-review-workflow skill by running the same review prompt against isolated before and after skill snapshots, collecting both outputs, and producing a…

    45k GitHub stars~1k tokensUpdated today
    Knowledge ManagementAuto-check passed
  • Obsidian CLI

    Atmosphere/atmosphere

    Interact with Obsidian vaults using the Obsidian CLI to read, create, search, and manage notes, tasks, properties, and more.

    3.8k GitHub starsUsed in 13 repos~795 tokens
    Knowledge ManagementAuto-check passed
  • Esm Cjs Risk Scan

    logseq/logseq

    Scan Logseq ClojureScript Node/Electron targets for npm module loading risks, especially ESM-only packages that may fail when loaded through js/require or shadow-cljs require-based shims.

    45k GitHub stars~3.3k tokensUpdated today
    Knowledge ManagementAuto-check passed
  • Capture Conversation

    outline/outline

    Save the current conversation, a decision, or a set of notes as a document in an Outline collection; use when the user wants to keep what was discussed in their knowledge base.

    41k GitHub stars~474 tokensUpdated today
    Knowledge ManagementAuto-check passed
  • Karpathy LLM Wiki

    Astro-Han/karpathy-llm-wiki

    A skill your agent uses when building or maintaining a personal LLM-powered knowledge base.

    2.4k GitHub stars~3.6k tokensUpdated 2 mo ago
    Knowledge ManagementAuto-check passed

Questions about Auto Skill

What does Auto Skill do?

在使用者明確同意的前提下,管理 Agent 任務的本地知識庫與跨技能經驗索引;當需要讀取、整理、驗證或回寫可重用的工作經驗時使用。絕不自動修改 IDE 或 Agent 全域規則。. Auto Skill is an agent skill from Toolsai/auto-skill.

When should I use Auto Skill?

Auto Skill fits situations like: knowledge Management work in your project.

How do I install Auto Skill in Claude Code?

Run `npx skills add Toolsai/auto-skill --skill auto-skill -a claude-code`. Or copy the skill folder (the Toolsai/auto-skill repository) into .claude/skills/auto-skill in your project. Claude Code loads it when a task matches its description.

How do I install Auto Skill in Codex?

Run `npx skills add Toolsai/auto-skill --skill auto-skill -a codex`. Or copy the skill folder (the Toolsai/auto-skill repository) into .agents/skills/auto-skill in your project. Codex loads it when a task matches its description.

Can I use Auto Skill in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add Toolsai/auto-skill --skill auto-skill -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/auto-skill, .gemini/skills/auto-skill, .github/skills/auto-skill and .opencode/skills/auto-skill in your project.

What does Auto Skill need to run?

Going by SKILL.md and its folder, Auto Skill needs the command-line tools its instructions call (npm). Our summary lists: Node.js.

Does Auto Skill access the network?

SKILL.md contains no URLs. Its commands use npm, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Auto Skill safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Auto Skill use?

Auto Skill is published under the MIT licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Auto Skill use?

About 991 tokens (SKILL.md is roughly 4k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Auto Skill?

Skills that share tags, products or a category with Auto Skill: Baoyu URL To Markdown (sdyckjq-lab/llm-wiki-skill, 2.5k stars), Logseq Review Workflow Eval (logseq/logseq, 45k stars), Obsidian CLI (Atmosphere/atmosphere, 3.8k stars) and Esm Cjs Risk Scan (logseq/logseq, 45k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Auto Skill?

Toolsai (a GitHub user) maintains it in Toolsai/auto-skill, which has 193 GitHub stars. The repository was last updated on August 17, 2026.

Source: Toolsai/auto-skill on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.