소스 정보
- 저장소
- diegosouzapw/awesome-omni-skill
- 최근 소스 활동
- 2026년 2월 28일 04:03
- 감지된 SKILL.md 언어
- 영어
- 스타
- 50
- 포크
- 19
설치 방법
기본적으로 소스를 먼저 확인하는 Prompt가 선택됩니다. 직접 명령으로 전환하거나 로컬 사본을 다운로드할 수도 있습니다.
소스 파일 검토
설치 여부를 결정하기 전에 SKILL.md와 SkillsMP에 표시된 보조 파일을 읽어 보세요.
메뉴
기본적으로 소스를 먼저 확인하는 Prompt가 선택됩니다. 직접 명령으로 전환하거나 로컬 사본을 다운로드할 수도 있습니다.
설치 여부를 결정하기 전에 SKILL.md와 SkillsMP에 표시된 보조 파일을 읽어 보세요.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
직접 명령은 검토 Prompt를 거치지 않습니다. 실행하기 전에 소스를 확인하세요.
npx skills add https://github.com/diegosouzapw/awesome-omni-skill --skill implement-js명령은 한 줄로 유지됩니다. 복사하기 전에 가로로 스크롤해 전체 내용을 확인하세요.
로컬 사본을 원하시나요? SkillsMP에서 현재 제공할 수 있는 파일을 다운로드하세요.
Token-efficient tracking for AI orchestration. CLI-first for status updates (~50 tokens), agent fallback for complex ops (~1KB). Use when: updating task status, querying blockers, creating progress files, validating phases.
AshAi extension guidelines for integrating AI capabilities with Ash Framework. Use when implementing vectorization/embeddings, exposing Ash actions as LLM tools, creating prompt-backed actions, or setting up MCP servers. Covers semantic search, LangChain integration, and structured outputs.
This skill should be used when solving hard questions, complex architectural problems, or debugging issues that benefit from GPT-5 Pro or GPT-5.1 thinking models with large file context. Use when standard Claude analysis needs deeper reasoning or extended context windows.
SOC 직업 분류 기준
SKILL.md 표시 중
| name | implement-js |
| description | Implement JavaScript code based on user requirements. |
| globs | ["*.js","package.json"] |
When implementing JavaScript code, follow these guidelines.
Use tabs and double quotes for strings
Make sure to follow style guide from eslint or/and editorconfig if present in the project
Use Common.JS syntax for imports and exports
I do not like to use this format const { recordEvent } = require("./event-log");, instead, do this const recordEvent = require("./event-log").recordEvent;
Write exports directly in module.exports = {};
Expose constants at the top of the file, after imports, for easier configuration
Use JSDoc when function parameters are ambiguous (e.g., options objects with multiple properties)
Prefer if/else over ternaries for non-trivial logic - readability over cleverness
Make sure to always create folder for each module, and create index.js inside, it helps with organization and reduces chance of conflict
I like to keep code for talking to other services in src/services folder
I like to keep helper code that is specification specific in src/utils folder. If it is not logically part of application, just something that is needed to make things work, it should go to utils
I create entrypoints into project in bin/ folder (e.g., bin/pull.js, bin/sync.js)
All entrypoints must be:
forever (e.g., forever bin/pull.js)package.json scripts (e.g., "pull": "forever bin/pull.js")Procfile for herokuish (e.g., pull: yarn pull)-watch variant using nodemon for development (e.g., "pull-watch": "nodemon bin/pull.js")If nodemon is not in package.json, install it as dev dependency (yarn add -D nodemon)
Scripts run in heroku/herokuish environment
Use Promise instead of async/await for asynchronous operations
Chain promises sequentially with .then() rather than nesting:
return doFirst()
.then(() => {
return doSecond();
})
.then(() => {
return doThird();
});
Avoid using ES6+ features that are not widely supported in Node.js environments
I like functional JS, map/reduce/foreach/filter should be preferred over for loops
I like to use wrappers for repeating code, like rate limiting, cachine, etc.
Never use var, always const/let. Unless you really need to change value, use const
Never change parameters of functions, always create new variables if you need to change something
Handle errors at the last possible moment to avoid losing error data when rebuilding. Catch earlier only when it makes sense for the logic
Try to encapsulate complex logic in modules and expose logic so that final code looks almost like pseudocode
Name modules based off their functionality, not how they are implemented (for example, module for logging stuff to mongo should be called logger, not mongo)
Where it makes sense, if there is array getter, also implement getter that returns map, calling original getter and transforming result to map
Services should be defined as objects with methods, not separate functions
Use _client for the HTTP client (private by convention), do not export it
Reference the service object by name inside methods
Example:
const got = require("got-verbose");
const config = require("../../config");
const myService = {
_client: got.extend({
prefixUrl: "https://api.example.com",
headers: {
authorization: "Bearer " + config.get("MY_API_KEY")
}
}),
getItems: () => {
return myService._client.get("items").then((response) => {
return JSON.parse(response.body).items;
});
},
getItemsMap: () => {
return myService.getItems().then((items) => {
const map = {};
items.forEach((item) => {
map[item.id] = item;
});
return map;
});
}
};
. = myService;
Always use yarn instead of npm, unless there is a package-lock.json present (then use npm)
Never guess npm package versions — use yarn add <package> (or yarn add -D <package> for dev dependencies) and let yarn resolve the correct version. Do not manually edit dependency versions in package.json.
Make sure there is .gitignore file present
node_modulesLib I like to use for rate limiting (how many calls of method should happen at the same time) is queue-promised
Install it, package name is queue-promised
Import method wrapper from queue-promised (const wrapper = require("queue-promised").wrapper;)
Use it like this const limitedFunction = wrapper(originalFunction, 5); where 5 is how many calls can happen at the same time
I like to use uuid package for generating uuids. I like to use v4 method from it. and I import it like this: const uuid = require("uuid").v4;
I like to use node-cron for cron jobs. I import it like this: const cron = require("node-cron");
I like to use got-verbose, which is a wrapper around got library for HTTP requests. It exposes identical API as got, but has built-in logging and error handling. I import it like this: const got = require("got-verbose");
Note: got-verbose does not work with streams. For streaming, use got directly (const got = require("got");)
Avoid using .json() method of got-verbose and json option, instead use .body and parse JSON manually, it does have some unexpected behavior
I like to use forever for running Node.js applications. Use it in the start script: "start": "forever src/index.js"
Add note to CLAUDE.md that this skill should be always used
Suggest changes to this skill, when I suggest code style changes
If you notice a pattern in my instructions, suggest adding it to this skill
Try to keep CLAUDE.md in sync with project changes
If you are writing a pseudocode (calling method or API endpoint you are not sure exists), add TODO comment so I know I should fix it
It is usual for me to use herokuish build like this
FROM gliderlabs/herokuish:latest
COPY . /app
WORKDIR /app
RUN /bin/herokuish buildpack build
ENV PORT 3000
EXPOSE 3000
When I tell you it is cleanup time, follow this checklist:
/implement-js
update docs (README.md)
update your docs (CLAUDE.md)
update missing jsdoc...
check tests
check linter