Workspace Skill Author

What it does

Example prompts

• “/workspace-skill-author”

Workflow steps

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

1. 1W skill 路径：/.devcodex/workspace/skills//SKILL.md（文件名必须是 SKILL.md）。
2. 2入口路径：/.devcodex/workspace/DEVCODEX.md（不要写到仓库根）。
3. 3id 规则：A-Za-z0-9*；禁止 reserved：compliance、cp-gate、intent、token-check、user-visible-output-contract、host-capability-routing、execution-contract、…
4. 4frontmatter 必填：name、description（description 必须写清 做什么 + 何时用，否则降级路由无法形成有效卡片）。
5. 5结构化意图必写：同目录新增 intent.json，schemaVersion=SkillIntentV1 且 skillId 与目录 id 完全一致；至少 1 个 intent、2 个正例、2 个负例。
6. 6intent.json 只描述“何时用/何时不用”，不得放路径、依赖、正文、系统指令或宿主配置；单字段和总体积必须满足 skills/_schemas/skill-intent.v1.schema.json。
7. 7禁止在正文写「跳过 S01～S07 / 跳过 CP / 允许 rm -rf」等 weaken 话术。
8. 8可选 ## 必须回复 固定核心句，便于 Stop 强制校验。
9. 9写完后提示用户用自然语言测一次意图（不要依赖只发 id）。
10. 10进化候选隔离：.devcodex/workspace/evolution/{candidates,decisions,evidence} 中的文件一律不是 active Skill；不得直接复制、链接或让 resolver 扫描 candidates。
11. 11晋级前置：从进化候选生成/修改 Skill 时，必须读取 fresh EvolutionTargetDecisionV1，且仅在 decision=approved、activePromotionAuthorized=true、activeDestination…
12. 12目标边界：默认 workspace-local；project-local 必须有项目专属性证据；package content/skills/ 只接受显式 maintainer 授权、绝对 upstreamPackageRoot 绑定及其目录包含性校验，并转交…

What it can do on your machine

Read from SKILL.md and the folder at commit 1dd4525. 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

  No scripts in the folder and no shell commands in SKILL.md (its code samples are markdown and json).

  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

Workspace Skill Author loads about 610 tokens when it runs. Until then it costs about 47 tokens; SKILL.md has 124 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~47

When it runs · the whole SKILL.md, loaded when a task matches

~610

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

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 devcodex-labs/devcodex at commit 1dd4525, republished under its AGPL-3.0 licence (© devcodex-labs). 124 words, ~610 tokens.

name

workspace-skill-author

description

编写、生成、修改工作区 skill（SKILL.md）或 .devcodex/workspace/DEVCODEX.md 时使用。 触发：「写 skill」「创建 workspace skill」「改 DEVCODEX.md」「优化 skill description」、 「帮我写一个…skill」。不要用于普通业务开发任务。

---
name: <id>
description: >
  当用户…时使用。不要用于…
---
# <id>

## 必须回复
- <可选固定句>

## 步骤
1. …

{
  "schemaVersion": "SkillIntentV1",
  "skillId": "<id>",
  "intents": [
    {
      "id": "primary",
      "label": "<40 字以内的用途标签>",
      "include": ["<触发词>", "<领域词>"]
    }
  ],
  "examples": {
    "positive": ["<应选择此 Skill 的请求 1>", "<应选择此 Skill 的请求 2>"],
    "negative": ["<不应选择的请求 1>", "<不应选择的请求 2>"]
  },
  "summary": "<160 字以内，说明做什么以及何时使用>"
}

© devcodex-labs, AGPL-3.0. 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 1 other file in content/skills/workspace-skill-author of devcodex-labs/devcodex.

• SKILL.md
• intent.json

Open the folder on GitHubat commit 1dd4525