Agent skill

Skill Authoring

by uvwt in uvwt/agentdock

创建、设计、修改、重构和验证 AgentDock Skill 时使用;负责 Agent Skills 兼容的 SKILL.md、可移植核心、引用、辅助脚本、测试、安全边界和本地真实验证。

Apache-2.0Auto-check: notesAgent Workflows

Install Skill Authoring

skills CLI
$ npx skills add uvwt/agentdock --skill skill-authoring -a claude-code

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

GitHub CLI
$ gh skill install uvwt/agentdock skill-authoring --agent claude-code

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

Manual copy
$ git clone --depth 1 https://github.com/uvwt/agentdock.git skills-src && mkdir -p .claude/skills && cp -r skills-src/core-skills/skill-authoring .claude/skills/skill-authoring && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

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

Facts

Skill name
skill-authoring
GitHub stars
1.2k
Used in
1 other repo
Token cost
~1.5k tokens
SKILL.md length
450 words
Files
4 (incl. references)
Skills in repo
3
Repo updated
First seen
Licence
Apache-2.0

At a glance

创建、设计、修改、重构和验证 AgentDock Skill 时使用;负责 Agent Skills 兼容的 SKILL.md、可移植核心、引用、辅助脚本、测试、安全边界和本地真实验证。

  • Works in 7 steps: 先定义模型何时应该选择该 Skill,再写正文。 → Skill 核心应兼容 Agent Skills 的文档模型,不把… → 最小 Skill 只有一个 SKILL.md;只有确有需要时才增加… → …
  • Tasks that involve Skill authoring
  • SKILL.md covers 核心原则, 目录, SKILL.md Frontmatter and AgentDock 路由索引, plus 7 more sections
  • Runs Python scripts from its folder; calls python3

What it does

Skill Authoring is an agent skill from uvwt/agentdock. 创建、设计、修改、重构和验证 AgentDock Skill 时使用;负责 Agent Skills 兼容的 SKILL.md、可移植核心、引用、辅助脚本、测试、安全边界和本地真实验证。

Its SKILL.md is about 1.5k tokens, which your agent loads only when the skill is triggered. The skill folder holds 5 other files, including reference files (for example `references/skill-package-spec.md`, `run.py` and `tests/test_run.py`).

It sits in Agent Workflows, covering Skill authoring. It works with Model Context Protocol. The repository describes itself as: Secure MCP runtime for AI agents to operate local machines, servers, and containers with multi-device orchestration. The licence is Apache-2.0.

When your agent uses it

  • Tasks that involve Skill authoring

Example prompts

  • “/skill-authoring”

Requirements

  • Python 3
  • A credential in EXAMPLE_API_KEY

Workflow steps

7 steps, taken from the first numbered list in SKILL.md.

  1. 先定义模型何时应该选择该 Skill,再写正文。
  2. Skill 核心应兼容 Agent Skills 的文档模型,不把 AgentDock 私有运行时当作通用契约。
  3. 最小 Skill 只有一个 SKILL.md;只有确有需要时才增加 references/、脚本或测试。
  4. 包内资源使用相对路径,业务环境变量由宿主注入当前进程。
  5. 秘密、缓存、会话、数据库、下载结果和设备私有状态不得进入 Skill 包。
  6. 修改后必须真实 lint、测试,并在需要时安装到 AgentDock 验证当前内容。
  7. 不创建或恢复 Skill 自有版本选择、激活、回滚或历史 revision 机制。

What it can do on your machine

Read from SKILL.md and the folder at commit a19d133. 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

    Ships script files (Python), which the agent can run.

    Shell commands in SKILL.md call:

    • python3

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

  • Network

    No URLs in SKILL.md.

    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

Skill Authoring loads about 1.5k tokens when it runs, and up to ~3.1k if it reads all its reference files. Until then it costs about 27 tokens; SKILL.md has 450 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~27
When it runs · the whole SKILL.md, loaded when a task matches
~1.5k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~3.1k

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: notes

The automated check noted patterns worth knowing about, such as sudo or a known installer.

  • NoteMentions a .env fileSKILL.md:201
    - 无 `.env`、数据库、截图、下载结果、`__pycache__`、`node_modules` 等运行产物;

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 uvwt/agentdock at commit a19d133, republished under its Apache-2.0 licence (© uvwt). 450 words, ~1,519 tokens.

Download SKILL.mdSave it as .claude/skills/skill-authoring/SKILL.md (or your agent's skills folder). This skill also uses 3 other files; get the full folder from GitHub.
name
skill-authoring
description
创建、设计、修改、重构和验证 AgentDock Skill 时使用;负责 Agent Skills 兼容的 SKILL.md、可移植核心、引用、辅助脚本、测试、安全边界和本地真实验证。

Skill Authoring

用于创建或维护 AgentDock Skill。Skill 的本体是模型可读取的说明文档;真实操作仍由命令、文件、浏览器、MCP 等工具完成。

AgentDock 不为 Skill 建立独立版本生命周期。Skill 的当前内容由来源与内容摘要识别;同名 managed Skill 更新时直接原子替换当前内容,不保留可选历史 revision,不提供 activate 或 rollback。

核心原则

  1. 先定义模型何时应该选择该 Skill,再写正文。
  2. Skill 核心应兼容 Agent Skills 的文档模型,不把 AgentDock 私有运行时当作通用契约。
  3. 最小 Skill 只有一个 SKILL.md;只有确有需要时才增加 references/、脚本或测试。
  4. 包内资源使用相对路径,业务环境变量由宿主注入当前进程。
  5. 秘密、缓存、会话、数据库、下载结果和设备私有状态不得进入 Skill 包。
  6. 修改后必须真实 lint、测试,并在需要时安装到 AgentDock 验证当前内容。
  7. 不创建或恢复 Skill 自有版本选择、激活、回滚或历史 revision 机制。

完整规范见 references/skill-package-spec.md。

目录

普通第一方和社区 Skill 默认放在独立 Skill 仓库:

text
skills/<skill-name>/
└── SKILL.md

按需扩展:

text
skills/<skill-name>/
├── SKILL.md
├── references/
├── scripts/
├── run.py
└── tests/

只有必须随 AgentDock runtime 一起交付的核心 Skill 放在:

text
core-skills/<skill-name>/

不要创建空目录。

SKILL.md Frontmatter

AgentDock 使用 Agent Skills 风格的字段:

yaml
---
name: example-skill
description: 清楚说明何时使用、解决什么问题
license: Apache-2.0
compatibility: Requires Python 3.11 or later.
metadata:
  owner: example-team
allowed-tools: exec_command
---

# Example Skill

AgentDock 真正依赖并严格校验的只有:

  • name 必填,长度 1–64,只允许小写 ASCII 字母、数字和 -,不能以 - 开头/结尾,也不能包含连续 --;
  • description 必填,最长 1024 个 Unicode 字符,并能让模型稳定判断何时使用;重要的 Use when、Do not use 和相邻 Skill 边界条件不要依赖正文补充,因为模型会先用 description 做候选路由;
  • Markdown 正文必须非空;
  • 其他 frontmatter(包括 license、compatibility、metadata、allowed-tools、version 以及第三方扩展字段)由作者生态定义,AgentDock 原样保留但不作为安装/运行前提;
  • 不设计 AgentDock 私有的 version、active_version、revision 或 rollback 契约。

AgentDock 路由索引

AgentDock 会先暴露轻量 description 索引,再按需读取完整 SKILL.md。不同来源的索引预算不同:

  • standalone managed 与 Plugin-owned Skill:agentdock_context.skills 返回 trim 后的完整 description;
  • workspace Skill:workspace_context.workspace_skills 返回 trim 后的完整 description,但最多列出 50 项;
  • shared/common Skill:agentdock_context.common_skills 面向数量不可控的 ~/.agents/skills,最多列出 50 项,并把每项 description 限制在 120 bytes。

因此 authoring 时应把 description 视为路由契约,而不是正文摘要的随意前缀。AgentDock 不提供可配置的 description 截断上限;如果未来 Skill 总量显著增长,应通过索引总预算或检索式路由解决,而不是静默裁掉每个已管理 Skill 的 description 后半段。

可移植核心

目标 Skill 默认只假设:

  • 当前进程能读取 Skill 包内容;
  • 包内资源使用相对路径;
  • 环境变量来自当前进程;
  • 宿主提供命令、文件、浏览器或远端工具能力。

有根目录脚本时,通用示例写成:

bash
printf '%s' '{"skill_action":"status"}' | python3 run.py

不要把以下内容作为核心运行前提:

  • ~/.agentdock/skills/<name>;
  • AGENTDOCK_HOME、AGENTDOCK_SKILL_DIR 等 AgentDock 私有路径变量;
  • skill_ref、exec_command 或 skill://;
  • 固定用户绝对路径;
  • 主动读取或 source AgentDock 私有环境文件。

AgentDock 专属说明可以放在独立的“AgentDock 适配/验证”章节;删除该章节后,核心流程仍应成立。

环境变量

需要配置时,在正文中明确声明变量名、类型、必填性和用途:

markdown
## 环境变量

| 变量 | 类型 | 必填 | 说明 |
|---|---|---:|---|
| EXAMPLE_BASE_URL | config | 是 | 服务地址 |
| EXAMPLE_API_KEY | secret | 是 | API Key |

Skill 只声明变量,不保存真实值。辅助脚本只从当前进程环境读取。

AgentDock 本地验证时:

  • skill_manage env_list 查看变量名称与配置状态;
  • skill_manage env_set 写入用户明确提供的值;
  • skill_manage env_unset 删除指定变量;
  • exec_command skill_ref=<host-issued-ref> 运行时只把 managed Skill 环境注入对应子进程。

环境值不写入 AgentDock 主进程或系统全局环境。

需要持久可变状态时,不要写 Skill 包目录。AgentDock 对 standalone managed 与 Plugin-owned Skill 的 exec_command 都会提供独立运行时保留变量 SKILL_DATA_DIR:

  • standalone managed 指向 ~/.agentdock/data/skills/<name>/;
  • Plugin-owned 指向 ~/.agentdock/data/skills/.plugin/<plugin>/<skill>/,并额外获得 Plugin 共享兼容目录 PLUGIN_DATA_DIR=~/.agentdock/data/plugins/<plugin>/;
  • 数据目录只在对应 Skill / Plugin 运行时真正需要时创建;
  • Unix 权限收紧为 0700,Windows 使用当前用户私有 ACL;
  • Windows 原生命令收到 Host 路径;WSL 命令收到已转换的 Linux 路径;
  • skill_manage env_set、宿主 env mapping 和 request.env 都不能覆盖运行时保留变量;
  • Plugin-owned Skill 的用户环境隔离在 ~/.agentdock/env/skill/plugin/<plugin>/<skill>.env;standalone 仍使用 ~/.agentdock/env/skill/<name>.env;
  • shared/workspace 候选不会得到 SKILL_DATA_DIR 或 PLUGIN_DATA_DIR。

SKILL_DATA_DIR / PLUGIN_DATA_DIR 是 AgentDock 可选适配,不是 Agent Skills 通用前提。可移植 Skill 不应把它们列为用户必填配置;需要状态目录的辅助脚本可以在检测到它们时优先使用,并在其他宿主下采用自己明确声明的可移植策略。

Show full SKILL.md (168 more words)Show less

引用与辅助脚本

正文引用包内文件时使用相对路径,例如:

text
references/api.md

辅助脚本应:

  • 通过 stdin 接收 JSON 对象;
  • 顶层动作字段使用 skill_action;
  • secret 只从环境读取;
  • 输出结构化 JSON;
  • 错误返回稳定 code 和可读 message;
  • 默认状态检查只读;
  • 破坏性动作要求明确确认;
  • 不回显 secret;
  • 用相对路径或脚本目录定位包内只读资源;
  • 可变缓存、数据库、会话和下载结果不得写回 Skill 包;在 AgentDock managed 运行时优先写入 SKILL_DATA_DIR。

Authoring lint

本 Skill 的 run.py 提供:

  • status:返回 lint_version 与规则数量;
  • lint:检查可移植性和明显宿主绑定。

示例:

json
{
  "skill_action": "lint",
  "source": "/path/to/skills/example-skill"
}

硬错误包括:

  • 硬编码 AgentDock managed Skill 安装目录;
  • 依赖 AgentDock 私有目录变量;
  • 主动读取 AgentDock 环境文件;
  • 固定用户绝对路径。

AgentDock 专属 skill_ref、exec_command、skill:// 出现在目标 SKILL.md 中会作为 warning,要求确认它们只存在于可选宿主适配说明。

安全与质量检查

提交前至少检查:

  • 无真实密码、Token、Cookie、私钥和认证缓存;
  • 无 .env、数据库、截图、下载结果、__pycache__、node_modules 等运行产物;
  • 无符号链接、路径逃逸、固定用户绝对路径;
  • 网络、写入、删除、上传、权限变化和依赖安装均有明确说明;
  • 未恢复 agentdock.yaml、skill_run、skill_env_manage、旧式 operation/entrypoint 清单或统一 Skill Runtime;
  • 不把安装历史、激活版本或回滚机制重新写回 Skill。

AgentDock 本地验证

AgentDock 的 managed Skill 当前布局是:

text
~/.agentdock/skills/<name>/
  SKILL.md
  ...

这个目录是宿主实现细节,不应硬编码进目标 Skill。

建议验证流程:

  1. 检查目录、Frontmatter 和正文;
  2. 检查禁止文件、符号链接和真实 secret;
  3. 对辅助脚本做语法检查和自带测试;
  4. 运行 skill-authoring lint,确认 portable=true;
  5. 使用 skill_manage install 从本地目录、压缩包或 HTTPS 来源安装;
  6. 记录返回的 content_digest;相同内容重复安装应为 no-op;
  7. 通过 agentdock_context 找到该候选,并保存它返回的 skill_ref 与 file;
  8. 用 read_file 读取宿主返回的 file,不要自己按名称重建 URI;
  9. 有辅助脚本时,用同一候选的 skill_ref 调用 exec_command 做只读检查;
  10. 修改内容后再次安装,确认当前内容被原子替换且环境/数据未被清空。

managed、shared、workspace 中同名 Skill 是不同候选。验证时始终使用宿主返回的精确 skill_ref,不得靠裸名称重新解析。

完成标准

只有同时满足以下条件才算完成:

  • description 可稳定触发且职责边界清楚;
  • Frontmatter 与 Agent Skills 文档模型兼容;
  • 可移植核心不依赖 AgentDock 私有目录或工具参数;
  • 包内脚本可从 Skill 根目录按相对路径运行;
  • lint、测试和安全检查通过;
  • managed 安装时有有效 content_digest,重复相同内容为 no-op;
  • agentdock_context 返回正确来源、skill_ref 与 file;
  • 当前内容和代表性只读行为已真实验证;
  • 没有 Skill 自有版本选择、激活、回滚或历史 revision 语义。

© uvwt, Apache-2.0. 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 3 other files (references) in core-skills/skill-authoring of uvwt/agentdock.

  • SKILL.md
  • references/skill-package-spec.md
  • run.py
  • tests/test_run.py

Open the folder on GitHubat commit a19d133

Used in 1 other repository

We found 1 copy of this SKILL.md (exact, near-identical or edited) in other folders, from 1 other GitHub owner. This page covers the copy in uvwt/agentdock, which our catalogue first saw on October 7, 2026.

Compare with similar skills

Skill Authoring 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.

Skill Authoring compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Skill Authoring this skilluvwt/agentdock1.2k1 repos~1.5kAutomated safety check: NotesApache-2.0
Skill Seekers Builderyusufkaraaslan/Skill_Seekers15k—~760Automated safety check: PassMIT
Mistral Vibe Plugin Creatormistralai/mistral-vibe5.1k—~3.1kAutomated safety check: PassApache-2.0
Scheduler Skill Creatoragenvoy/Agenvoy564—~3kAutomated safety check: PassAGPL-3.0
Workflow Schema Tuningbreaking-brake/cc-wf-studio5.4k—~1.3kAutomated safety check: PassCustom licence
Skill Creatoragenvoy/Agenvoy564—~3kAutomated safety check: PassApache-2.0

Similar skills

  • Skill Seekers Builder

    yusufkaraaslan/Skill_Seekers

    Detects the type of a knowledge source and uses the Skill Seekers MCP tools to turn docs, repos, PDFs or videos into packaged AI skills.

    15k GitHub stars~760 tokensUpdated 8 days ago
    Agent WorkflowsAuto-check passed
  • Mistral Vibe Plugin Creator

    mistralai/mistral-vibe

    Official

    Shows how to build a Vibe plugin package in the Agent Plugins 1.0 format, with a plugin.json manifest and optional skills, MCP servers, hooks and other components.

    5.1k GitHub stars~3.1k tokensUpdated yesterday
    Agent WorkflowsAuto-check passed
  • 建立並排程定時觸發的 skill。所有新增定時/週期任務、提醒、排程通知的請求必須走此 skill,禁止直接呼叫 schedules(mode=write)(那是 skill 已存在時的時間綁定工具,不該作為新建排程的入口)。

    564 GitHub stars~3k tokensUpdated yesterday
    Agent WorkflowsAuto-check passed
  • Workflow Schema Tuning

    breaking-brake/cc-wf-studio

    Guides edits to cc-wf-studio's workflow schema so AI agents generate better workflows, treating schema text as prompt engineering rather than validation.

    5.4k GitHub stars~1.3k tokensUpdated 2 days ago
    Agent WorkflowsAuto-check passed
  • Skill Creator

    agenvoy/Agenvoy

    Create, edit, improve, or audit AgentSkills. An agent skill from agenvoy/Agenvoy.

    564 GitHub stars~3k tokensUpdated yesterday
    Agent WorkflowsAuto-check passed
  • Skill Creator

    swarmclawai/swarmclaw

    Create, edit, improve, or audit skills for SwarmClaw agents.

    688 GitHub stars~1.4k tokensUpdated 3 mo ago
    Agent WorkflowsAuto-check passed

More from uvwt/agentdock

  • Agentdock User Guide

    uvwt/agentdock

    当用户询问 AgentDock 是什么、如何使用、配置在哪里、不同平台或安装方式怎样修改配置并生效、如何重启或验证配置、如何发现并配置 Codex/Claude/Grok 等 Coding Agent 的 ACP,以及常见运行问题时使用;覆盖 macOS Desktop、Windows Desktop、Linux 服务、Docker 和直接运行二进制,不用于源码开发与贡献流程。

    1.2k GitHub stars~1.6k tokensUpdated 2 days ago
    Auto-check passed
  • Plugin Import

    uvwt/agentdock

    当用户要安装或更新来自 Git、GitHub、外部插件市场或其他远程来源的 Plugin 时使用;负责把远程来源固定并取得到本地,再交给 pluginmanage 自动识别 Portable/OpenAI/Claude 格式、审核和安装。

    1.2k GitHub stars~839 tokensUpdated 2 days ago
    Auto-check passed

Categories

Questions about Skill Authoring

What does Skill Authoring do?

创建、设计、修改、重构和验证 AgentDock Skill 时使用;负责 Agent Skills 兼容的 SKILL.md、可移植核心、引用、辅助脚本、测试、安全边界和本地真实验证。. Skill Authoring is an agent skill from uvwt/agentdock.

When should I use Skill Authoring?

Skill Authoring fits situations like: tasks that involve Skill authoring.

How do I install Skill Authoring in Claude Code?

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

How do I install Skill Authoring in Codex?

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

Can I use Skill Authoring 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 uvwt/agentdock --skill skill-authoring -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/skill-authoring, .gemini/skills/skill-authoring, .github/skills/skill-authoring and .opencode/skills/skill-authoring in your project.

What does Skill Authoring need to run?

Going by SKILL.md and its folder, Skill Authoring needs Python for the scripts in its folder and the command-line tools its instructions call (python3). Our summary lists: Python 3; A credential in EXAMPLE_API_KEY.

Does Skill Authoring access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Skill Authoring safe to install?

Our automated static check of SKILL.md found notes only (mentions a .env file), nothing it rates as a warning. It is not a guarantee. Review the folder before installing.

What licence does Skill Authoring use?

Skill Authoring is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Skill Authoring use?

About 1.5k tokens (SKILL.md is roughly 6.1k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 1.6k tokens, read only when the agent opens those files.

What are the alternatives to Skill Authoring?

Skills that share tags, products or a category with Skill Authoring: Skill Seekers Builder (yusufkaraaslan/Skill_Seekers, 15k stars), Mistral Vibe Plugin Creator (mistralai/mistral-vibe, 5.1k stars), Scheduler Skill Creator (agenvoy/Agenvoy, 564 stars) and Workflow Schema Tuning (breaking-brake/cc-wf-studio, 5.4k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Skill Authoring?

uvwt (a GitHub user) maintains it in uvwt/agentdock, which has 1,197 GitHub stars. The repository holds 3 skills in this directory. The repository was last updated on October 7, 2026.

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