AgentStack
SKILL verified MIT Self-run

Skill Authoring

skill-fldicoahkiin-code-alchemist-skill-authoring · by Fldicoahkiin

Create, validate, and maintain Claude Code Skills. Use when writing a new skill, reviewing skill structure, or debugging skill-related issues.

No reviews yet
0 installs
22 views
0.0% view→install

Install

$ agentstack add skill-fldicoahkiin-code-alchemist-skill-authoring

✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README — it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-fldicoahkiin-code-alchemist-skill-authoring)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
4mo ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming — see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps — measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Skill Authoring? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Skill Authoring Guide

编写符合 Agent Skills 规范的 Claude Code Skill。

Use This Skill When

  • 创建新的 Claude Code Skill
  • 审查现有 Skill 的结构和规范性
  • 调试 Skill 验证失败的问题
  • 理解 Skill 文件组织和最佳实践

Skill 结构

一个标准的 Skill 目录结构:

.agents/skills//
├── SKILL.md              # 主文件(必需)
├── skill.lock.json       # 锁定文件(推荐)
├── evals/
│   └── evals.json        # 测试用例(推荐)
├── references/           # 参考资料(可选)
│   └── spec.md
├── scripts/              # 辅助脚本(可选)
│   └── validate.sh
└── templates/            # 模板文件(可选)
    └── snippet.md

SKILL.md 格式

Frontmatter(必需)

---
name: skill-name                          # 小写字母、数字、连字符
description: '简短描述,说明用途和触发时机。'  # 1-1024 字符,需引号
license: MIT                              # 开源许可证
metadata:
  author: Your Name
  version: 1.0.0
  tags: "tag1, tag2, tag3"
  repository: https://github.com/...
---

关键规则

  • name:只能使用小写字母、数字、连字符
  • description:必须加引号,长度 1-1024 字符,清晰说明何时使用该 skill
  • description 是第一段 frontmatter 的 description,不要被示例代码中的 frontmatter 干扰

Body 内容

示例结构:

# Skill 标题

## When to Use / 何时使用

  • 触发条件 1
  • 触发条件 2

## 核心工作流

  1. 步骤一
  2. 步骤二
  3. 步骤三

## 规则 / Patterns

### 分类一

  • 规则 1
  • 规则 2

### 分类二 ``bash # 代码示例 ``

## Anti-Patterns(禁止做的事)

  • 不要做 X
  • 避免 Y

## Scope(适用范围)

Apply to:

  • path/pattern/**/*

Do not over-apply to:

  • excluded/pattern/**/*

验证 Skill

验证当前 Skill(自验证)

在 skill 目录下运行:

bash scripts/validate_skill.sh

验证其他 Skill

验证指定路径的 skill:

bash .agents/skills/skill-authoring/scripts/validate_skill.sh path/to/your-skill

重要:如果不提供路径参数,脚本只会验证它自己(skill-authoring),而不是你新建的 skill。

验证项

  • Frontmatter 格式正确
  • Name 字段合法(小写字母、数字、连字符;与目录名一致)
  • Description 长度 1-1024 字符且带引号
  • License 和 metadata 字段完整
  • H1 标题存在
  • Body 行数不超过 500(推荐)
  • evals.json 存在且合法

常见错误

| 错误 | 原因 | 修复 | |------|------|------| | Description 长度异常 | 脚本读取了示例代码中的 frontmatter | 使用只提取第一段 frontmatter 的 awk 命令 | | Name 包含非法字符 | 使用了大写字母或下划线 | 改为小写和连字符 | | 缺少 H1 标题 | 没有 # Title | 添加一级标题 | | evals.json 格式错误 | JSON 语法错误 | 用 python3 -m json.tool 检查 |

skill.lock.json 格式

{
  "name": "skill-name",
  "version": "1.0.0",
  "lockfileVersion": 1,
  "skill": {
    "name": "skill-name",
    "version": "1.0.0",
    "entry": "SKILL.md",
    "files": [
      "SKILL.md",
      "scripts/helper.sh"
    ],
    "dependencies": {
      "git": ">=2.0.0"
    },
    "checksums": {
      "SKILL.md": "sha256...",
      "scripts/helper.sh": "sha256..."
    }
  },
  "generated": "2026-03-31T00:00:00Z"
}

evals.json 格式

{
  "skill_name": "skill-name",
  "evals": [
    {
      "id": 1,
      "prompt": "测试 prompt",
      "expected_output": "预期输出描述",
      "assertions": ["检查点 1", "检查点 2"]
    }
  ]
}

最佳实践

Do

  • 保持 body 在 500 行以内
  • 使用具体的文件路径模式
  • 提供清晰的代码示例
  • 包含 When to Use 章节
  • 声明 Anti-Patterns

Don't

  • 在 frontmatter 中放代码示例
  • 使用过长的 description
  • 遗漏 H1 标题
  • 创建没有测试用例的 skill

调试技巧

  1. 验证脚本报告 description 长度错误

``bash # 手动计算真实 description 长度 awk '/^---$/ {found++} found==1 && /^description:/ {sub(/^description:[[:space:]]*/, ""); print; exit}' SKILL.md | sed "s/^[\\\"']*//;s/[\\\"']*$//" | wc -c ``

  1. 检查 frontmatter 解析问题

``bash # 确认只提取了第一段 frontmatter awk '/^---$/ {found++} found<=2' SKILL.md ``

  1. 生成 checksum

``bash sha256sum SKILL.md ``


参考: https://agentskills.io/specification

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet — be the first.

Versions

  • v0.1.0 Imported from the upstream source.