SuperAPI AI Directory
Partner中文AI newsFavorites
Skillsxzskill

xzskill

Office & productivity

View repository

Updated 10-18 07:25

About this skill

Historical version and status

This is the complete historical guide declaring the name xzskill in orziz/odai, pinned to commit 4a694b0cc90bd48ecc3828af4beaf576cbbe2fe7 dated 2026-04-09 UTC. Observed 2026-10-18 UTC. The commit date is not a claimed first publication date for the skill.

The original catalog name, ID, repository and installation command are retained instead of being replaced with a later related entry. Historical workflows, host conventions and installation methods may have changed; current installability of the legacy entry is unverified. The source is reference material here. No installation, modification, review, synchronization, file cleanup or agent-dispatch instruction from it was executed. The source record retains the original Chinese; the English view translates prose while preserving code and source links.

Complete historical guide

name: xzskill description: Based on the standard skills in the skills directory, generate the corresponding manual installation versions and sync the README

xzskill Now executed via the in-repo script scripts/xzskill.js .

Goal

After the user passes in one or more skill names, based on 工程根目录/skills/<skill-name>/ generate the corresponding manual installation files, and perform minimal-scope synchronization on README.md .

Input constraints

  1. Accepts one or more skill names, for example:review-sslb harness-sslb。
  2. If no name is passed in, or any name cannot be unambiguously resolved to a skill name, prompt the user to re-enter.
  3. Only recognize the 工程根目录/skills/<skill-name>/ directory, not from .agents The directory takes the installation source.
  4. Do not consider dependencies and do not introduce other skill names.
  5. If <skill-name> is xzskill, report an error directly and stop, to avoid self-referential generation.

Execution method

  1. The Claude command layer only provides a thin wrapper.
  2. The actual generation logic is uniformly completed by 工程根目录/scripts/xzskill.js .
  3. The script is the sole execution truth; do not let the model manually generate target files according to rules anymore.

Script responsibilities

  1. Check whether 工程根目录/skills/<skill-name>/SKILL.md exists; if it does not exist, report an error and stop.
  2. Read 工程根目录/skills/<skill-name>/SKILL.md，use it as the main body source.
  3. If references/、assets/ or scripts/exists in the same directory, treat these directories as auxiliary resources of the current skill and sync them together to each manual installation target directory.
  4. .claude/commands/<skill-name>.md existing wrapper information (such as the supplementary notes after allowed-tools、argument-hint、用户输入 ) as the Claude-side source of truth, and preserve it during generation.
  5. Do not read other skill directories, do not scan unrelated directories in the repository, and do not refer to other skills as templates.
  6. When generating the target file, by default adopt the "fixed template + direct body copy" strategy:Do not rewrite the source file body, do not reorganize, and do not stylistically rewrite.
  7. If the body references relative paths under references/、assets/、scripts/ , when generating the manual installation file, rewrite these paths to relative paths usable in the target directory.
  8. First parse the source file frontmatter:
    • Read name
    • Read description
  9. Then take the main text after the frontmatter of the source file as the sole body payload; keep the paragraph order, heading hierarchy, code blocks, and wording of the main text unchanged.
  10. Generate or overwrite the following target files directly based on the source file content:
  • .claude/commands/<skill-name>.md
  • .github/skills/<skill-name>/SKILL.md
  • .trae/skills/<skill-name>.md
  • .trae/rules/<skill-name>.md
  1. If an auxiliary resource directory exists, synchronize the directory of the same name at the following location:
  • .claude/commands/<skill-name>/
  • .github/skills/<skill-name>/
  • .trae/skills/<skill-name>/
  • .trae/rules/<skill-name>/
  1. The generation rules for each target file are fixed as follows, with no on-the-spot judgment:
  • .claude/commands/<skill-name>.md: retain name、description, and preserve as much as possible the existing allowed-tools、argument-hint and 用户输入 in that file, then append explanatory notes; after the frontmatter, always append a blank line,用户输入：、$ARGUMENTS, a blank line, and then concatenate the retained section and the source file main text
  • .github/skills/<skill-name>/SKILL.md: retain name、description after the frontmatter, then directly concatenate the source file main text; place auxiliary resources and SKILL.md in the same skill directory, and keep the relative paths of resources in the main text in the same form as in the source file
  • .trae/skills/<skill-name>.md: retain name、description after the frontmatter, then directly concatenate the source file main text; do not rewrite additionally
  • .trae/rules/<skill-name>.md: retain name、description frontmatter, then directly concatenate the source file body; do not additionally rewrite
  1. When updating README.md , handle only the "currently provided" list:
  • If 工程根目录/skills/<skill-name>/SKILL.mddoes not exist in the list, append one line
  • If it already exists, do not add it again
  • Do not modify other sections of the README, and do not rewrite the entire text
  1. If multiple skills are passed in, complete all validation first, then process them one by one, to avoid synchronizing only half.
  2. Throughout, handle only the currently specified skill, and stop after completion.

Strict restrictions

  1. Do not modify any content under .agents .
  2. Do not scan or operate on unrelated skills.
  3. Do not invent new file formats, and do not extend the installation system.
  4. Do not search the entire repository for files just to "refer to existing formats".
  5. If the source skill does not exist, do not perform any writes.
  6. Except for 工程根目录/skills/<skill-name>/, the current .claude/commands/<skill-name>.md , and README.md , do not read any other skill directories.
  7. Do not rewrite the source skill body just to "look more like the target platform"; the body must always be copied verbatim from the source file, with no changes except for the fixed header, the Claude-side preserved section, and rewriting of resource relative paths.
  8. Do not reorder body sections, do not merge paragraphs, and do not add or remove rules on your own.
  9. Do not infer platform differences on your own, and do not design extra fields on your own; anything the source file does not state and the rules do not fix must not be added.
  10. If the generation rules are already fixed, execute according to the fixed template; do not repeatedly compare "whether it could look even more like the target platform".
  11. When writing to the target file, any occurrence of $ARGUMENTSin the rules must be treated as a plain-text placeholder: do not expand it, do not substitute it, and do not reference the currently passed-in parameters.

Failure output

  • If any source file does not exist:错误：工程根目录/skills/<skill-name>/SKILL.md 不存在，无法生成手动安装版本。
  • If passed xzskill：错误：xzskill 不支持生成自身的手动安装版本，请改为处理其他 skill。

Successful output

On success, briefly state only:

  1. which source files were processed;
  2. which target files were updated for each;
  3. whether the README was appended to, updated, or already existed;
  4. synchronization was completed within the minimal necessary scope.

Pinned source, license and attribution

Exact-name original; same-version source directory; same-version license. Original documents by orzi! under MIT. Original text is retained unchanged; the status notes and English translation are editorial additions.

MIT License

Copyright (c) 2026 orzi!

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.