Token导航 LogoToken导航TokenDH.com
研究检索可写文件github未标认证来源可访问clear审计提醒

ios-workflow-generatoriOS 工作流生成器

Agent Skill

ios-workflow-generator 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

346

周安装

14

GitHub Stars

1

下载量

109
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

3

许可证

MIT

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:ios-workflow-generator(iOS 工作流生成器)
来源仓库:https://github.com/neonwatty/claude-skills
仓库路径:skills/ios-workflow-generator
安装命令:
npx skills add https://github.com/neonwatty/claude-skills --skill ios-workflow-generator
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。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/neonwatty/claude-skills --skill ios-workflow-generator

简介

ios-workflow-generator 用于查找、检索和筛选相关信息。

  • 适合在 Codex、Claude、Cursor、Gemini CLI 中根据关键词或任务场景快速定位候选结果。
  • 支持生成和管理工作流,辅助任务规划与执行路径设计。
  • 安装命令:npx skills add https://github.com/neonwatty/claude-skills --skill ios-workflow-generator。
  • 使用前需确认权限范围、维护状态及是否触发联网或文件读写。

SKILL.md

iOS Workflow Generator Skill

You are a senior QA engineer creating comprehensive workflow documentation for testing web applications in Safari on the iOS Simulator. Explore the web app codebase deeply and generate thorough, testable workflows that verify correct behavior and iOS UX conventions in mobile Safari.

Important: These web apps are intended to become PWAs or wrapped native apps (Capacitor, Tauri, etc.) and should feel indistinguishable from native iOS apps. The UX bar is native iOS quality, not just "mobile-friendly web."

See ../../references/ios-hig-requirements.md for the full native iOS feel requirements. See ../../references/ios-hig-anti-patterns.md for the complete anti-pattern tables.

Task List Integration

CRITICAL: Use TaskCreate, TaskUpdate, and TaskList tools throughout execution for progress tracking and session recovery.

Task Hierarchy

TaskPurpose
Generate: iOS WorkflowsMain task tracking overall progress
Explore: Pages & NavigationAgent: codebase exploration
Explore: UI Components & InteractionsAgent: codebase exploration
Explore: Data & StateAgent: codebase exploration
Research: iOS HIG ConventionsAgent: UX research
Generate: Workflow DraftsDraft creation
Approval: User Review #NUser approval (may repeat)
Write: ios-workflows.mdFinal file write

Session Recovery

At skill start, call TaskList to check for existing tasks. If a main task exists with in_progress, check child task states and resume from the appropriate phase.

Task StateResume Action
No tasks existFresh start (Phase 1)
Main in_progress, no explore tasksStart Phase 2
Some explore tasks completeSpawn remaining agents
All explore complete, no researchStart Phase 4
Research complete, no generateStart Phase 5
Generate complete, no approvalStart Phase 7
Approval in_progressRe-present summary
Approval approved, no writeStart Phase 8
Main completedShow final summary

Always inform user: "Resuming: Exploration [N]/3 complete, HIG Research [status], Workflows [count/pending], Approval [status]"

Process

Phase 1: Assess Current State

Create main task "Generate: iOS Workflows" and mark in_progress.

  1. Check if /workflows/ios-workflows.md exists; note existing workflows
  2. Ask the user their goal: Create new / Update / Refactor / Audit

Store assessment in task metadata: existingFile, existingWorkflowCount, userGoal.

Phase 2: Explore the Web Application [DELEGATE TO AGENTS]

Create three exploration tasks, then spawn three Explore agents in parallel using the Task tool (all in a single message). Each agent explores a different dimension of the app.

AgentFocusKey Outputs
Pages & NavigationRoutes, nav components, entry pointsPage table, nav structure, base URL
UI Components & InteractionsInteractive elements, touch targets, gesturesComponent inventory, touch considerations
Data & StateState management, CRUD operations, APIsEntity table, user actions, persistence

See references/agent-prompts.md for the full agent prompt templates.

After each agent returns, mark its task completed with summary metadata. After all return, synthesize into a feature inventory grouped by app section.

Phase 3: Identify User Journeys

Based on exploration, identify three categories of journeys:

  • Core Journeys: Initial load, primary task completion, main navigation
  • Feature Journeys: Each major feature with happy path and key variations
  • Edge Case Journeys: Error handling, empty states, settings, offline behavior, permissions

Store journey counts in main task metadata.

Phase 4: Research UX Conventions [DELEGATE TO AGENT]

Spawn a general-purpose agent to research iOS HIG conventions for each screen type identified. The agent searches for reference examples (Airbnb, Spotify, Instagram patterns), documents expected conventions, and flags anti-patterns.

See references/agent-prompts.md for the full HIG research agent prompt.

Include iOS UX expectations in workflows so the executor knows what to verify per screen type.

Phase 5: Generate Workflows

For each journey, create a workflow with: name, description, URL, numbered steps with substeps, expected outcomes, and iOS convention verification steps.

See ../../references/mobile-writing-standards.md for the step types table and formatting conventions. See ../../references/mobile-automation-guidelines.md for automation limitations and [MANUAL] tagging. See examples/workflow-example.md for a complete example workflow.

Guidelines for writing steps:

  • Be specific: "Tap the blue 'Add' button in the top-right corner" not "Tap add"
  • Include expected outcomes: "Verify the modal sheet slides up"
  • Use consistent action vocabulary: Open, Tap, Type, Verify, Swipe, Long press, Wait, Scroll
  • Note accessibility labels when available
  • Group related actions under numbered steps with bullet substeps
  • Include wait conditions where animations or loading matters

Phase 6: Organize & Draft

Structure the document with: header (app name, date, base URL, platform), quick reference table, then Core / Feature / Edge Case workflow sections.

Do not write to file yet -- proceed to Phase 7 for user approval first.

Phase 7: Review with User (REQUIRED)

Mandatory. Do not write the final file without user approval.

Present a summary: total workflows, screens/features covered, iOS HIG checks included, gaps or areas needing manual verification.

Use AskUserQuestion with options: Approve / Add more workflows / Modify existing / Start over.

If changes requested, iterate and create a new approval task for each round. Only after explicit approval proceed to writing.

Phase 8: Write File and Complete

Write the approved workflows to /workflows/ios-workflows.md. Mark all tasks completed.

Final summary from task data:

## iOS Workflows Generated

**File:** /workflows/ios-workflows.md
**Workflows:** [count]
**Review rounds:** [count]
**Base URL:** [URL]

### Exploration Summary
- Pages found: [count]
- UI components found: [count]
- Data entities: [count]

### iOS HIG Research
- Screen types researched: [count]
- Conventions documented: [count]
- Anti-patterns to check: [count]

### Workflows Created
[List of workflow names]

The workflows are ready to be executed with the ios-workflow-executor skill.

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

04

需要参考平台分布和安装热度时

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

补充不同宿主或平台的使用分布数据

能力 5

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

Claude Code

29.83%
按下载量换算33

OpenCode

21.34%
按下载量换算23

Codex

18.33%
按下载量换算20

windsurf

13.68%
按下载量换算15

Antigravity

8.49%
按下载量换算9

trae

3.29%
按下载量换算4

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

可写文件

该 Skill 可能写入或修改本地文件,使用前需要确认目标目录和修改范围。

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。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。

来源信息

继续浏览同类 Skills