Token导航 LogoToken导航TokenDH.com
研究检索需要联网github未标认证来源可访问许可证需确认审计通过

workflow-designer工作流程设计师

Agent Skill

workflow-designer 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。

总安装

432

周安装

18

GitHub Stars

公开资料未说明

下载量

144
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:workflow-designer(工作流程设计师)
来源仓库:https://github.com/falak-vhora/workflow-designer-skill
仓库路径:skills/workflow-designer
安装命令:
npx skills add https://github.com/falak-vhora/workflow-designer-skill --skill workflow-designer
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/falak-vhora/workflow-designer-skill --skill workflow-designer

简介

用于查找、检索和筛选相关信息,支持工作流程设计与规划。

  • 适合在关键词或任务场景驱动下快速定位候选方案与来源线索。
  • 需结合原始 README 验证具体实现,注意可能的联网与文件操作。
  • 安装前建议确认权限范围和维护状态,防止非预期系统交互。
  • workflow-designer 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Workflow Designer

This skill helps the agent design clear, detailed, implementation-ready workflows for engineering tasks, with an emphasis on code-focused work (features, refactors, debugging, reviews, etc.).

The primary output is a checklist-style workflow that someone could follow end-to-end.

When to Use This Skill

The agent should apply this skill whenever:

  • The user explicitly asks to “design a workflow”, “workflow designer”, or similar.
  • The user asks for a step-by-step implementation plan (e.g. “give me a plan to implement X”).
  • The user asks for a process around technical work (e.g. “design a deployment process”, “review process”, “incident workflow”).
  • The task involves non-trivial multi-step engineering work where a structured process will help.

This skill is primarily for:

  • Code-centric workflows: implementing features, refactors, migrations, debugging, testing.
  • Engineering processes: deployment, release, code review, incident handling, onboarding technical tasks.

Output Format

Always produce a checklist-first workflow, optimized for someone to execute step by step.

Required Structure

Follow this structure unless the user specifies another:

## Context
- **Goal**: [...]
- **Scope**: [...]
- **Assumptions**: [...]

## High-Level Phases
- **Phase 1**: [...]
- **Phase 2**: [...]
- **Phase 3**: [...]

## Detailed Workflow Checklist
- [ ] **Phase 1: ...**
  - [ ] Step 1: ...
  - [ ] Step 2: ...
- [ ] **Phase 2: ...**
  - [ ] Step 1: ...
  - [ ] Step 2: ...

## Risks and Validation
- **Risks / pitfalls**:
  - ...
- **Validation / completion criteria**:
  - ...

Level of Detail

  • Default to detailed and exhaustive within reason:

- Break work into small, actionable steps that could be checked off. - Make implicit steps explicit when they are easy to forget (e.g. “run tests”, “update docs”).

  • Avoid unnecessary prose; prefer concise, actionable bullets.

If the scope is huge (e.g. “design complete architecture for a bank”), first:

  • Briefly note that the scope is broad.
  • Propose a phased workflow (e.g. discovery → design → implementation → rollout).
  • Optionally ask the user what phase to detail first, but still provide an initial high-level workflow.

Workflow Design Process (How the Agent Should Think)

When invoked, the agent should follow this internal process before writing the final workflow:

  1. Clarify the goal (mentally or briefly in text)

- What is the primary outcome? (e.g. “Feature X shipped to production”, “Refactor Y complete with no regressions”). - Is this mainly about implementation, refactor, migration, debugging, process, or other?

  1. Identify constraints and context

- Check for any constraints mentioned by the user: deadlines, tech stack, environments, approvals, testing requirements. - If constraints are missing but critical, either: - Infer sensible defaults, and - Optionally call them out in the Assumptions section.

  1. Define 3–6 high-level phases Examples:

- For a feature: *Discovery → Design → Implementation → Testing → Review → Release* - For a refactor: *Analysis → Design → Incremental changes → Testing → Cleanup & docs* - For a process: *Intake → Triage → Execution → Validation → Closure*

  1. Expand phases into checklists

- For each phase, create 3–10 concrete steps. - Prefer verbs at the start: “Analyze…”, “Implement…”, “Write…”, “Run…”, “Verify…”. - Include critical cross-cutting steps: testing, observability, documentation, communication.

  1. Add risks and validation

- List key pitfalls (e.g. data loss, downtime, regressions). - Provide explicit done criteria and validation checks (tests, monitoring, reviews).

  1. Tailor to the user’s context

- If the user’s stack, tools, or conventions are known from the conversation, reference them (e.g. “update README.md”, “add Jest tests”, “update CI pipeline”).

Patterns and Examples

Example: Implementing a New Feature

User request: “Design a workflow to implement a new user profile page in our web app.”

The agent should respond in this style (shortened here for brevity):

## Context
- **Goal**: Implement a new user profile page in the web app and release it safely to production.
- **Scope**: Backend API, frontend UI, tests, and deployment.
- **Assumptions**: Existing auth system; CI/CD pipeline in place.

## High-Level Phases
- **Phase 1**: Clarify requirements and UX
- **Phase 2**: Design API and data model changes
- **Phase 3**: Implement backend + frontend
- **Phase 4**: Testing and QA
- **Phase 5**: Launch and follow-up

## Detailed Workflow Checklist
- [ ] **Phase 1: Clarify requirements and UX**
  - [ ] Capture required fields and interactions for the profile page.
  - [ ] Identify permissions (who can view/edit which fields).
  - [ ] Agree on UX wireframes or rough layout.

- [ ] **Phase 2: Design API and data model changes**
  - [ ] Review existing user model and endpoints.
  - [ ] Design any new fields and validation rules.
  - [ ] Define API contract (request/response shapes).

... (and so on for other phases)

## Risks and Validation
- **Risks / pitfalls**:
  - Inconsistent data between old and new profile fields.
  - Leaking sensitive information on the profile page.
- **Validation / completion criteria**:
  - All acceptance criteria are met and covered by automated tests.
  - No P0 errors in logs after release window.

Example: Refactor Workflow

For refactors, emphasize safety and incremental change:

  • Phases like: *Baseline & safety nets → Refactor in small steps → Keep behavior identical → Clean up → Monitor*.
  • Checklist items for:

- Capturing current behavior, adding missing tests. - Refactoring in thin slices. - Running tests and static analysis after each logical change.

Example: Debugging Workflow

For debugging requests:

  • Phases like: *Reproduce → Narrow scope → Form hypotheses → Test systematically → Fix → Prevent regression*.
  • Checklist items for:

- Capturing failing cases and logs. - Binary search / isolation strategies. - Adding regression tests once fixed.

Adapting to User Requests

  • If the user specifies a preferred structure (e.g. “only numbered steps”, “no risks section”), respect their format while keeping the spirit of this skill (clear, actionable steps).
  • If the user wants a lighter or heavier workflow, adjust detail accordingly but keep the Context → Phases → Checklist → Validation pattern unless explicitly told otherwise.

Summary of Key Behaviors

When this skill is active, the agent should:

  • Think like a senior engineer designing a process: prioritize safety, clarity, and maintainability.
  • Produce detailed checklists that someone else can follow without extra context.
  • Call out assumptions, risks, and validation steps explicitly.
  • Adapt the workflow to the specific task, stack, and constraints the user mentions.

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

39.34%
按下载量换算57

Claude

29.55%
按下载量换算43

Cursor

20.01%
按下载量换算29

Gemini CLI

9.08%
按下载量换算13

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

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