Token导航 LogoToken导航

Agent Skill · 研究检索

the-standard-core(标准核心)

the-standard-core 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

the-standard-core 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为只读;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 14,932。

总安装 8,197周安装 388GitHub Stars 17下载 4,273核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:the-standard-core(标准核心)
来源仓库:https://github.com/hassanhabib/the-standard-skills
仓库路径:skills/the-standard-core
安装命令:
npx skills add https://github.com/hassanhabib/the-standard-skills --skill 'The Standard Core'
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/hassanhabib/the-standard-skills --skill 'The Standard Core'

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

the-standard-core 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中快速定位候选结果。适用于根据关键词、任务场景或来源线索进行信息检索与筛选的场景。通过 npx skills add 命令从指定 GitHub 仓库安装并使用。

安装前建议确认权限范围、维护状态,以及是否会触发联网或文件读写操作。可结合原始 README 进一步核验具体用法和功能边界。

来源仓库内路径:skills/the-standard-core

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/hassanhabib/the-standard-skills --skill 'The S…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

the-standard-core 怎么安装?

在终端执行 npx skills add https://github.com/hassanhabib/the-standard-skills --skill 'The Standard Co…。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

the-standard-core 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

the-standard-core 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

the-standard-core 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

the-standard-core 用的人多吗?

公开数据是累计安装 8,197 次,近一周安装 388 次,下载 4,273 次,GitHub 17 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

the-standard-core 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。
  • 标注为只读权限,试用阶段不容易误改本地文件。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。
  • 只想先试试效果、不希望它改动本地文件的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
the-standard-core

标准核心 · 本页

8,19717只读未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

The Standard Core

What this skill is

This skill is the governing layer for every other Standard skill. It defines the theory, operating model, and core values that must shape all design, implementation, testing, review, and planning decisions.

The Standard is technology-agnostic. Apply it regardless of language, framework, runtime, or hosting model. Use C# and.NET examples only as materializations of the theory, not as the theory itself.

Explicit coverage map

This skill explicitly covers the foundational material of The Standard:

  • 1. Introduction
  • 0.0 The Theory

- 0.0.0 Introduction - 0.0.1 Finding Answers - 0.0.2 Tri-Nature - 0.0.2.0 Purpose - 0.0.2.1 Dependency - 0.0.2.2 Exposure - 0.0.3 Everything is Connected - 0.0.4 Fractal Pattern - 0.0.5 Systems Design & Architecture - 0.0.6 Conclusion

  • 0.1 Purposing, Modeling, and Simulation

- 0.1.0 Introduction - 0.1.1 Purposing - 0.1.1.0 Observation - 0.1.1.1 Articulation - 0.1.1.2 Solutioning - 0.1.2 Modeling - 0.1.2.0 Model Types - 0.1.2.0.0 Data Carrier Models - 0.1.2.0.0.0 Primary Models - 0.1.2.0.0.1 Secondary Models - 0.1.2.0.0.2 Relational Models - 0.1.2.0.0.3 Hybrid Models - 0.1.2.0.1 Operational Models - 0.1.2.0.1.0 Integration Models (Brokers) - 0.1.2.0.1.1 Processing Models (Services) - 0.1.2.0.1.2 Exposure Models (Exposers) - 0.1.2.0.2 Configuration Models - 0.1.3 Simulation - 0.1.4 Summary

  • 0.2 Principles

- 0.2.0 People-First - 0.2.0.0 Simplicity - 0.2.0.0.0 Excessive Inheritance - 0.2.0.0.1 Entanglement - 0.2.0.0.1.0 Horizontal Entanglement - 0.2.0.0.1.1 Vertical Entanglement - 0.2.0.0.2 Autonomous Components - 0.2.0.0.2.0 No Magic - 0.2.1 Rewritability - 0.2.2 Mono-Micro - 0.2.3 Level 0 - 0.2.4 Open Code - 0.2.5 Airplane Mode (Cloud-Foreign) - 0.2.6 No Toasters - 0.2.7 Pass Forward - 0.2.8 All-In/All-Out - 0.2.9 Readability over Optimization - 0.2.10 Last Day

When to use

Use this skill always when any software-related request is present. Use it before architecture, before coding, before testing, before planning, and before review. If another Standard skill applies, activate this one first and let it govern the rest.

Identity and theory

  1. Every system is governed by theory, whether explicit or implicit.
  2. The Standard is powered by the Tri-Nature theory.
  3. Every system must be understood through three lenses:

- Purpose - Dependency - Exposure

  1. Tri-Nature is fractal.

- The same pattern repeats at system level, sub-system level, service level, validation level, and exposure level.

  1. Everything is connected.

- Design every part with awareness that it will become someone else’s dependency or exposure.

Mandatory engineering sequence

Follow this order intentionally:

  1. Purposing
  2. Modeling
  3. Simulation

This order is mandatory at initiation. Iteration is allowed later, but do not skip purpose to jump into code.

Purposing rules

  1. Start with observation.

- Identify the real blocker, constraint, or unmet need.

  1. Articulate the problem clearly.

- Use words, diagrams, or figures when helpful. - A well-described problem is halfway solved.

  1. Include solutioning in the purpose.

- Do not treat the path to the goal as trivial. - A solution must also honor readability, configurability, longevity, optimization, and maintainability.

  1. Never cut corners to reach the goal.

- Reaching the goal the wrong way is a violation.

  1. Always keep purpose visible.

- If the purpose is unclear, stop and clarify before modeling or implementation.

Modeling rules

  1. Model only what the purpose requires.
  2. Do not attract irrelevant attributes into the model.
  3. Treat models as classes that express only the data required for the problem.
  4. Prefer the most generic valid name that still fits the problem scope.

Data carrier model rules

  1. Primary models are self-sufficient.

- They do not physically depend on another model to exist.

  1. Secondary models depend on primary models.

- They usually reference another model or are nested within one.

  1. Relational models connect two primary models.

- They exist to materialize many-to-many relationships. - They should hold references, not unrelated details.

  1. Hybrid models are allowed only when necessary.

- They mix relational behavior with additional details about the relationship. - Prefer purity first; use hybrid models only when the business flow truly requires it.

Operational model rules

  1. Integration models are brokers.
  2. Processing models are services.
  3. Exposure models are exposers.
  4. Configuration models are startup, composition, middleware, or entry-point models.

Simulation rules

  1. Simulation is how models interact.
  2. A model may act on another model.
  3. A model may be acted upon by another model.
  4. A model may act on itself.
  5. Functions, methods, and routines are simulation mechanisms.
  6. Keep simulation inside the purpose and the model boundaries.

Core principles and non-negotiables

People-First

  1. Build for both users and future maintainers.
  2. Favor human understanding over cleverness.
  3. Favor systems that mainstream engineers can own, evolve, and rewrite.

Simplicity

  1. Simplicity is mandatory, not optional.
  2. Simplicity creates rewritability.
  3. Simplicity favors modular monoliths and clear decomposition.
Excessive Inheritance
  1. Do not use more than one level of inheritance.
  2. More than one level is excessive and prohibited, except when vertical versioning of flows absolutely requires it.
Entanglement
  1. Avoid horizontal entanglement.

- No Utils. - No Commons. - No Helpers that pretend to simplify everything. - No shared models across independent flows. - No shared exceptions or shared validation rules across unrelated flows.

  1. Avoid vertical entanglement.

- No local base components unless they are native or external. - No local inheritance chains that create hidden coupling.

  1. Prefer duplication over cross-flow entanglement when duplication preserves autonomy.

Autonomous Components

  1. Every component should be self-sufficient.
  2. Every component owns its validations, tooling, and utilities in one of its dimensions.
  3. Components may depend on injected dependencies, not hidden helpers.
  4. Duplication is permitted when it preserves ownership and autonomy.

No Magic

  1. What you see is what you get.
  2. No hidden routines.
  3. No magical extensions that require chasing references.
  4. No runtime tricks that make the system hard to understand.
  5. Put validation, exception handling, tracing, security, localization, and behavior in plain sight.

Rewritability

  1. Every system must be easy to understand, modify, and fully rewrite.
  2. A Standard-compliant system should be forkable, clonable, buildable, and testable with minimal surprise.
  3. No hidden dependencies.
  4. No unknown prerequisites.
  5. No injected routines that obscure behavior.

Mono-Micro

  1. Build monoliths with a microservice mindset.
  2. Every flow should be autonomous enough to be extracted later.
  3. Modularize aggressively without splitting prematurely.

Level 0

  1. Code must be understandable by entry-level engineers.
  2. Level 0 understanding is the measure of success.
  3. If new engineers cannot follow the system, the system is too complex.

Open Code

  1. Prefer code, tooling, platforms, and patterns that are visible and accessible.
  2. Do not make source proprietary solely for personal or organizational gain.
  3. Exceptions exist only when safety, security, or contractual obligations require it.

Airplane Mode (Cloud-Foreign)

  1. The system should be runnable locally without mandatory cloud dependency.
  2. Develop tooling that bridges cloud resources to local stand-ins.
  3. Favor local testing, local development, and mocked external systems.

No Toasters

  1. Do not force Standard compliance via style cops or analyzers as the primary mechanism.
  2. Teach the Standard person-to-person.
  3. Favor conviction and understanding over coercion.
  4. AI-assisted coding (including “vibe coding”) is acceptable, provided a human remains actively involved and retains full responsibility for the final code.
  5. AI may be used for code reviews and suggestions, but the final authority for approval and merge decisions must always rest with a human.

Pass Forward

  1. The Standard is to be taught at no cost.
  2. Do not profiteer from teaching The Standard.
  3. Do not use The Standard to belittle others.
  4. Pass it forward freely.

All-In / All-Out

  1. The Standard must be embraced fully or rejected fully.
  2. Partial adoption is not standardization.
  3. Outdated partial adherence is not enough to claim a Standardized system.

Readability over Optimization

  1. When in doubt, choose readability.
  2. Unreadable “optimal” software is not truly optimum.

Last Day

  1. Work every day as if it might be your last day on the project.
  2. End each engineering day at a good stopping point.
  3. Ensure another engineer can pick up the work seamlessly the next day.
  4. Apply this to design, development, documentation, and automation.

Mandatory operating behavior for agents

  1. Never skip purpose.
  2. Never start with implementation when purpose or models are unclear.
  3. Never optimize at the expense of readability.
  4. Never recommend hidden shared abstractions as a first move.
  5. Never recommend analyzers or “toasters” as a substitute for understanding.
  6. Reject vague or chaotic designs.
  7. Prefer explicitness, autonomy, and rewritability.
  8. Use the language of The Standard consistently.
  9. If there is a conflict between generic convention and The Standard, The Standard wins.

平台分布

Codex

36.82%
按下载量换算1,573

Claude

31.88%
按下载量换算1,362

Cursor

18%
按下载量换算769

Gemini CLI

9.43%
按下载量换算403

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills