Token导航 LogoToken导航TokenDH.com
开发只读github未标认证来源可访问许可证需确认审计通过

advanced-engineer高级工程师

Agent Skill

advanced-engineer 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。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。

总安装

372

周安装

16

GitHub Stars

公开资料未说明

下载量

131
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

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:advanced-engineer(高级工程师)
来源仓库:https://github.com/benmix/skills
仓库路径:skills/advanced-engineer
安装命令:
npx skills add https://github.com/benmix/skills --skill advanced-engineer
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/benmix/skills --skill advanced-engineer

简介

advanced-engineer 强调工程化执行流程,要求先理解问题本质再修改代码,聚焦根因修复而非症状修补。

  • 适用于 Bug 定位、测试失败排查及集成故障分析等需要严谨证据链支持的场景。
  • 采用 Plan-Act-Reflect 循环机制,每次仅做最小化变更并验证预期效果,保持代码一致性。
  • 操作前务必确认对 GitHub 仓库的读写权限,避免触发未授权的文件修改或命令执行行为。
  • 适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

SKILL.md

Advanced Engineer

Use this skill when the task requires disciplined engineering execution rather than a quick patch.

Mission

  • Understand the problem before changing code.
  • Fix root causes, not symptoms.
  • Keep changes minimal and consistent with the surrounding code.
  • Verify the result with real evidence before claiming success.

Default Loop

Work in a strict Plan -> Act -> Reflect loop:

  1. Plan the next concrete step and what evidence it should produce.
  2. Act by reading code, running commands, or making the smallest justified change.
  3. Reflect on what changed, what was learned, and whether the hypothesis still holds.

Do not chain random actions together without updating the model of the problem.

Start Here

Keep In SKILL Context

  • Do not patch code you do not yet understand.
  • Prefer explicit hypotheses over intuition.
  • Increase investigation depth when an attempt fails.
  • Preserve existing behavior unless the task requires changing it.
  • Be explicit when verification is partial.

Reference Index

Read only the references needed for the current task.

references/01-engineering-fundamentals.md

Read for:

  • change scope discipline
  • avoiding over-engineering and premature abstractions
  • reading code before editing
  • boundary validation and secure-by-default decisions
  • what to do when an approach is blocked

references/02-planning-and-execution.md

Read for:

  • iterative planning
  • decomposition for larger tasks
  • parallel work or worker orchestration
  • completion protocol for delegated work

references/03-verification-and-review.md

Read for:

  • adversarial verification
  • writing a verification plan
  • review and simplification passes before declaring completion

references/04-safe-delivery.md

Read for:

  • destructive action thresholds
  • git safety
  • commit, PR, and hook discipline
  • when to confirm with the user before taking irreversible actions

references/05-debugging-and-root-cause-analysis.md

Read for:

  • Plan -> Act -> Reflect
  • investigation before modification
  • root-cause-first debugging
  • hypothesis-driven debugging
  • escalation strategy after failed attempts
  • verification pipeline and partial-verification rules
  • tool-first investigation and when to ask the user

Completion Criteria

Treat the task as complete only when:

  • the root cause is understood
  • the fix addresses that cause
  • the result is verified, or the verification limit is stated clearly
  • nearby regressions have been checked
  • the chosen references were actually applied to the task at hand

Do not report success based on a plausible diff alone.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

31.68%
按下载量换算42

Claude

31.21%
按下载量换算41

Cursor

20.33%
按下载量换算27

Gemini CLI

8.63%
按下载量换算11

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。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills