ryanallen

Create Skills

知識管理v0.1.2
作者:ryanallen更新於 2026-05-11共 3 個版本 · 最新版本釋出於 2026-03-10document-skills@clawhub_ryanallen/document-skills
知識庫問答資料整理檔案問答#Document

據 AlphaMoat 統計,截至 2026-10,Create Skills 累計下載 11,358 次、獲 12 個星標,按累計下載量在本次參與排名的 10,914 個 AI 技能中排名第 2,419 位,在「知識管理」分類(745 個)中排名第 163 位。

列表預設按累計下載量排序,也可按星標排序;詳情頁排名按累計下載量計算。

累計下載
1.1萬
星標
12
全站下載量排名
#2,419
全站前 23% · 「知識管理」前 22%
平台累計下載星標評論
ClawHub ↗60851-
SkillHub ↗1.1萬12—

各平台的計數口徑不同,統計時間不一致。

這個技能是做什麼的

整理輸入、輸出與過程,檢查前置資訊和目錄命名,並區分指令碼、參考資料與素材。除非使用者要求改行為,文件應描述現有行為;重新命名或移動技能須更新引用。

資料釋出於 2026-10-27

可以用來做什麼

  • 讓技能文件符合宿主約定

什麼時候用

  • 建立或修訂技能檔案時

示例提示詞

以下是根據技能文件生成的示例,不代表實際執行結果。

檢查技能結構與後設資料,提出文件修改,並指出重新命名會影響的引用。

月度資料趨勢

累計下載1.1萬
243872291.2萬2026-03 · 累計下載 30992026-05 · 累計下載 46602026-06 · 累計下載 51142026-07 · 累計下載 54962026-08 · 累計下載 58202026-09 · 累計下載 67202026-10 · 累計下載 1.1萬26-0326-0926-10
星標12
07132026-03 · 星標 12026-05 · 星標 62026-06 · 星標 112026-07 · 星標 112026-08 · 星標 122026-09 · 星標 122026-10 · 星標 1226-0326-0926-10

2026-09 起統計方法調整:同名技能按作者分別統計,並覆蓋全部技能。圖中虛線為調整前的月份,與之後的資料不直接比較,也不計算跨越調整的增長。

月份累計下載環比星標
2026-101.1萬—12
2026-096720—12
2026-085820+5.9%12
2026-075496+7.5%11
2026-065114+9.7%11
2026-054660—6
2026-033099—1

SKILL.md 原文

以下是作者釋出的原始檔案,按純文字顯示,本站未做修改或稽核;其中的連結不可點選,命令請核實後再執行。抓取於 2026-10-23。

13 KB · ClawHub v0.1.2
---
name: document-skills
description: Write or update a skill (SKILL.md and supporting files) to match host structure and best practices. Use when user says create skill, write skill, update SKILL.md, /document-skills.
disable-model-invocation: true
argument-hint: "[skill-path] [source]"
context: fork
agent: general-purpose
---

# Document Skills
展開全部(共 149 行)

Write or update a skill to fit the host's expected structure. Use the structure and checklist below.

## Inputs

1. **Target skill** – Path to the skill folder (e.g. the project skills directory `skills/example/`) or the skill name.
2. **Source** – Draft, user instructions, or notes to turn into the skill or merge in.

**If you don't give either:** Use the current or given context (e.g. the skill or path already under discussion).

## Output

SKILL.md updated (and supporting files if needed).

## Process

### 1. Skill layout

- One folder per skill. The main file is always `SKILL.md` (exact name, case-sensitive).
- **Folder name:** Must be **kebab-case** (lowercase, hyphens; no spaces, underscores, or capitals). The frontmatter `name` must match the folder name.
- **Where skills live:** In a project, typically a dedicated skills directory (e.g. `skills/<name>/`). The host discovers skills from this location even when it's nested (e.g. inside a package).
- **Optional subfolders only:** Use **scripts/** (executable code), **references/** (documentation loaded as needed), and **assets/** (templates, icons, etc.). Do not put README.md inside the skill folder; all documentation goes in SKILL.md or references/. Long reference material goes in `references/` and is linked from SKILL.md (progressive disclosure).
- **Length:** Keep SKILL.md under ~5,000 words. Long reference in `references/`, linked from SKILL.md.
- **Progressive disclosure:** Skills use three levels — (1) YAML frontmatter is always loaded so the assistant knows when to use the skill; (2) SKILL.md body is loaded when the skill is relevant; (3) files in references/ or assets/ are opened only as needed. This keeps token use down while keeping expertise available.
- **Composability:** Skills can load together. Write the skill so it works alongside others; don't assume it's the only one active.

### 2. SKILL.md format

#### Frontmatter (the YAML block between the first `---` lines)

Frontmatter configures when and how the skill runs. The table below follows the [official skills reference](https://code.claude.com/docs/en/skills.md#frontmatter-reference).

**name** and **description** are required. **name** must match the folder name (kebab-case). **description** should follow: **`[What it does] + [When to use it] + [Key capabilities]`** — under 1024 characters; no XML angle brackets (`<` `>`). Avoid vague lines like "Helps with projects"; include specific trigger phrases and, if relevant, file types (e.g. ".fig files", "PDF contract review").

| Field                      | Required    | Description                                                                                                                                           |
| :------------------------- | :---------- | :---------------------------------------------------------------------------------------------------------------------------------------------------- |
| `name`                     | Yes         | kebab-case only; must match skill folder name. Lowercase letters, numbers, and hyphens (max 64 characters).                                           |
| `description`              | Yes         | **`[What it does] + [When to use it] + [Key capabilities]`**. Include trigger phrases (e.g. "Use when user says save, /save"). Mention file types if relevant. Under 1024 characters. No `<` or `>`.        |
| `argument-hint`            | No          | Hint shown during autocomplete to indicate expected arguments. Example: `[issue-number]` or `[filename] [format]`.                                    |
| `disable-model-invocation` | No          | Set to `true` to disable automatic loading; use for workflows you want to trigger manually with `/name`. Default: `false`. |
| `user-invocable`           | No          | Set to `false` to hide from the `/` menu. Use for background knowledge users shouldn't invoke directly. Default: `true`.                              |
| `allowed-tools`            | No          | Tools the assistant can use without asking permission when this skill is active.                                                                             |
| `model`                    | No          | Model to use when this skill is active.                                                                                                               |
| `context`                  | No          | Set to `fork` to run in a forked subagent context.                                                                                                    |
| `agent`                    | No          | Which subagent type to use when `context: fork` is set.                                                                                               |
| `hooks`                    | No          | Hooks scoped to this skill's lifecycle. See [Hooks in skills and agents](/en/hooks#hooks-in-skills-and-agents) for configuration format.              |

**Optional:** `license` (e.g. MIT for open source), `compatibility` (1–500 chars: product, system deps, network), `metadata` (e.g. author, version, mcp-server). See [complete skills guide](https://resources.anthropic.com/hubfs/The-Complete-Guide-to-Building-Skill-for-Claude.pdf).

**Security:** No XML angle brackets in frontmatter. Avoid reserved names in the skill name (e.g. host or vendor names).

**YAML gotcha:** Avoid colons inside a value. They can be read as a new key. Use something like "Scope is" instead of "Scope:" in the text.

#### Content type

- **Reference content** – Conventions, patterns, style. The assistant applies it inline to your current work. No invocation control needed.
- **Task content** – Step-by-step instructions for a specific action (deploy, commit, code generation). Often invoked directly with `/skill-name`. Use `disable-model-invocation: true` so the skill is not auto-loaded.

#### Body sections (same order for every skill)

Use the same section order so skills are easy to scan:

1. **# Title** – Skill name as H1, then one short intro (what it does).
2. **## Inputs** – What the skill needs (user input, paths, options). Number the items. Use "None" or "Optional" when nothing is required.
3. **## Output** – What you get (a file, a behavior, a handoff). One short block.
4. **## Process** – How to do it. Numbered steps or subsections (e.g. "### 1. Step name"). Put all how-to and rules here. Be specific and actionable (e.g. "Run `python scripts/validate.py --input {filename}`" not "Validate the data"). Include error handling for common failures; reference bundled files (e.g. "Before queries, consult `references/api-patterns.md`") when relevant.
5. **## Examples** – Optional but recommended. Common scenarios: "User says X → Actions → Result."
6. **## Troubleshooting** – Optional but recommended. Error or symptom → Cause → Solution.
7. **## Reference** – Optional. Links to related skills or docs.

Skip a section only if it really doesn't apply. When unsure, include Inputs, Output, Process, and Reference; add Examples and Troubleshooting where they help. Match the layout of other skills in the repo.

#### Substitutions (in the body text)

From the [official docs](https://code.claude.com/docs/en/skills.md#available-string-substitutions):

- `$ARGUMENTS` – All arguments passed when invoking the skill. If not present in the content, arguments are appended as `ARGUMENTS: `.
- `$ARGUMENTS[N]` or `$N` – The Nth argument (0-based index).
- `${CLAUDE_SESSION_ID}` – The current session ID (host-defined).
- `${CLAUDE_SKILL_DIR}` – The directory containing the skill's `SKILL.md` file (host-defined).

For injecting shell output before the skill runs, see [Inject dynamic context](https://code.claude.com/docs/en/skills.md#inject-dynamic-context).

### 3. Checklist when writing or updating

1. **Frontmatter:** `description` follows **`[What it does] + [When to use it] + [Key capabilities]`** and is specific (include phrases users might say, e.g. "Use when user says save, /save"). `name` matches the skill. Task or side-effect skills: `disable-model-invocation: true` per official docs.
2. **Sections:** Inputs, Output, Process, then optionally Examples and Troubleshooting, then Reference. Same order as other skills.
3. **Length:** Under 500 lines; long reference in other files, linked from SKILL.md.
4. **Supporting files:** If used, say so in SKILL.md and when to load them. Keep SKILL.md focused; put long or detailed docs in `references/` and link (progressive disclosure).
5. **Arguments:** Use `$ARGUMENTS` or `$N` and optionally `argument-hint` if the skill takes input.
6. **Invocation:** Use `disable-model-invocation` and/or `user-invocable` so the skill runs when intended (see [Control who invokes a skill](https://code.claude.com/docs/en/skills.md#control-who-invokes-a-skill)).

### 4. Steps to run

1. Read the target skill path and any source.
2. Apply the checklist and structure above. Don't change behavior unless the user asked.
3. Write or update SKILL.md (and supporting files if needed). Only document what the skill does; don't add capabilities that aren't there.

## Examples

**User says:** "Create a skill for validating CSV uploads."

**Actions:** Create folder `skills/validate-csv/` (or the project's skill path; kebab-case). Add SKILL.md with required frontmatter (`name: validate-csv`, description using **`[What it does] + [When to use it] + [Key capabilities]`**, e.g. "Validate CSV uploads. Use when user says validate CSV, check upload, /validate-csv. Key capabilities: schema check, encoding detection."). Add Inputs, Output, Process, optional Reference. If the skill runs a script, add `scripts/validate.sh` and reference it in Process.

**Result:** New skill that triggers on the stated phrases; coordinator and flows updated if it is part of a flow.

**User says:** "Update the save skill to mention update-gitignore."

**Actions:** Open the save skill's SKILL.md (e.g. `skills/save/SKILL.md`). Add the new step or reference in Process. Update description if trigger phrases change. Sync coordinator and [references/coordinator-flows.md](../../agents/references/coordinator-flows.md) if the Save flow changes.

## Troubleshooting

**Invalid frontmatter / upload error.**  
Cause: YAML formatting (missing `---` delimiters, unclosed quotes, or colons in a value read as a new key).  
Solution: Ensure frontmatter is between two `---` lines. Avoid colons in description text; use "Use when" not "When:". No `<` or `>` in frontmatter.

**Skill doesn't trigger.**  
Cause: Description too vague or missing the full formula (what it does, when to use it, key capabilities) or trigger phrases.  
Solution: Use **`[What it does] + [When to use it] + [Key capabilities]`** and add specific "Use when user says X, Y, /skill-name". Same phrases should appear in coordinator flow table if the skill is part of a flow. **Debug:** Ask the assistant "When would you use the [skill name] skill?" and adjust the description from what’s missing.

**Skill triggers too often.**  
Cause: Description too broad or overlapping with other skills.  
Solution: Add negative triggers (e.g. "Do NOT use for simple data exploration; use data-viz skill instead"). Narrow scope (e.g. "PDF legal documents for contract review" instead of "Processes documents"). Clarify when not to use (e.g. "Use specifically for online payment workflows, not general financial queries").

**Wrong folder name.**  
Cause: Folder has spaces, underscores, or capitals.  
Solution: Rename to kebab-case. Update `name` in SKILL.md to match. Update all references (coordinator, agents, README, package.json, checklist script, [agents/references/](../../agents/references/)).

## After writing

- **Coordinator sync:** If you changed the description (e.g. "when to use" or key capabilities that match user requests), update the coordinator agent and the flow steps in [references/coordinator-flows.md](../../agents/references/coordinator-flows.md) if the skill is part of a flow, so the flow table and agent descriptions stay in sync.
- **Rename/move:** If a skill was **renamed or moved** (e.g. generate-figma → designer-figma), update all references: coordinator, agents, README, package.json, other skills that link to it, [verify-task checklist script](verify-task/scripts/checklist.ts), [agents/references/](../../agents/references/) (e.g. coordinator-flows.md if the skill appears in a flow), and .gitignore.

## Reference

[Official skills reference](https://code.claude.com/docs/en/skills.md) – frontmatter, control who invokes a skill, run in subagent, substitutions. [Coordinator](../../agents/coordinator.md) – Flow table and agent descriptions should follow **`[What it does] + [When to use it] + [Key capabilities]`** and match the phrases users say.

[Complete guide to building skills](https://resources.anthropic.com/hubfs/The-Complete-Guide-to-Building-Skill-for-Claude.pdf) – progressive disclosure, description formula, good/bad examples, optional frontmatter, triggering and troubleshooting, instructions best practices.

「知識管理」分類下的熱門技能

搜尋引擎知識管理

透過公開搜尋URL在16個搜尋引擎(7個國內直連+9個國際)中做跨引擎檢索、結果去重與多源交叉驗證,無需API Key,內建零依賴Python指令碼。Use when 使用者要求“搜一下/查資料/多找幾個來源/核實訊息/找公眾號文章/排查報錯/找論文或開源專案”時。輸入自然語言查詢詞,輸出帶來源連結與可信度標註的結構化結果。不適用於私有資料檢索、批次抓取、需要登入後可見的內容。

↓ 1075.9萬★ 147u_8ee34d77
ima-skills知識管理

透過 IMA OpenAPI 管理筆記與知識庫內容。

↓ 108.4萬★ 566u_1742680f
Summarize知識管理

把長文字、文件與網頁內容壓縮成要點和關鍵詞的摘要工具。

↓ 87.6萬★ 1048paudyyin
全網搜尋知識管理

透過“查一下、搜一下、搜尋、找一下”等自然語言檢索公開網頁,整理來源並交付橫版 PDF、Markdown 與標題連結表。支援使用者主動開啟、檢視和取消技能收藏,保留原報告取回。

↓ 50.9萬★ 2u_32d99f8a
web-tools-guide知識管理

規定如何為網頁請求選擇合適工具的必讀指引。

↓ 40萬★ 220user_ec205dbb
Aihot Skill Lite知識管理

AI HOT(aihot.virxact.com)中文 AI 資訊查詢 Skill,幫助使用者獲取今日 AI 圈動態、AI 日報、AI 熱點、最新 AI 資訊以及 OpenAI、Anthropic、Google 等公司的最新發布資訊。

↓ 32.1萬★ 92Khazix
全球12億文獻知識庫(8千萬中文期刊可下載)知識管理

全球12億文獻知識庫(8千萬中文期刊可下載)——透過 SmartLib 開放平台 API 提供中外文學術文獻檢索與下載能力,覆蓋 8000 萬篇授權中文期刊全文 + 12.28 億條全球文獻後設資料(期刊 7.19 億 / 專利 2.15 億 / 會議 7155 萬 / 學位論文 2473 萬 / 標準 268 萬)。 能力:中英文關鍵詞檢索、文獻詳情、中文期刊 PDF 全文下載、外文 OA 文獻十級渠道免費下載(不消耗配額)、智慧關鍵詞擴充套件、核心期刊優先排序、相關性重排、引文追溯、分類號檢索。 配額:首次使用自動註冊,免費 100 次檢索 + 10 次下載 / 月;耗盡自動彈出方案(體驗卡 / 個人版月 / 專業版月 / 單篇下載 / 下載包),可說「升級 / 充值」喚起;企業 / 機構定製聯絡 [email protected]。 觸發:使用者表達"查論文""找文獻""檢索學術""搜期刊""查專利""找標準""下論文""寫文獻綜述""找參考文獻""查 SCI/EI"等意圖時啟用;也適用"幫我找關於 XX 的論文""寫文獻綜述""找幾篇引用支撐論點"。英文:"find papers" "search literature" "write literature review" "find supporting citations"。 呼叫前必須先用 /consume 獲取 consume_token,再憑 token 調 /search(每次計費介面呼叫都需一次 /consume)。

↓ 30.3萬★ 189user_164f4c1f
Markdown Converter知識管理

用 uvx 直接呼叫 markitdown,把各類檔案轉成 Markdown。

↓ 20.4萬★ 256Peter Steinberger

資料月份:2026-10 · 下載/星標數聚合自公開技能註冊庫(ClawHub、SkillHub),按月更新。口徑說明見資料方法。