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

user-flow-mapping用户流量映射

Agent Skill

user-flow-mapping 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

4,125

周安装

126

GitHub Stars

67

下载量

1,497
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:user-flow-mapping(用户流量映射)
来源仓库:https://github.com/seb1n/awesome-ai-agent-skills
仓库路径:skills/user-flow-mapping
安装命令:
npx skills add https://github.com/seb1n/awesome-ai-agent-skills --skill 'User Flow Mapping'
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/seb1n/awesome-ai-agent-skills --skill 'User Flow Mapping'

简介

用于处理 GitHub 仓库、Issue、Pull Request 等协作信息,适合代码变更管理场景。

  • 适用于围绕仓库状态和代码变更进行整理和协作的场景。
  • 通过 npx skills add 命令从指定 GitHub 仓库安装使用。
  • 安装前需确认权限范围和维护状态,注意可能触发联网或命令执行。
  • user-flow-mapping 属于待分类类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

User Flow Mapping

This skill enables the agent to create detailed user flow diagrams that map every step, decision point, error state, and success path a user encounters while completing a task in a product. The agent produces three types of flows — task flows (single path, no decisions), user flows (multiple paths with decision branches), and wire flows (flows annotated with screen wireframes) — using Mermaid diagram syntax for portability. Each flow includes annotations for conversion metrics, drop-off risk points, and optimization opportunities.

Workflow

  1. Define the Flow Objective and Scope: Identify the specific user goal being mapped (e.g., "Complete a purchase," "Reset a password"). Determine the entry points — how the user arrives at the start of the flow (direct link, homepage navigation, email CTA, push notification). Establish the success criteria and the scope boundary so the diagram does not expand indefinitely.
  2. Identify All Steps and Decision Points: List every screen, action, and system response in sequence. Mark decision points where the user or the system branches (e.g., "Is the user logged in?" or "Did payment succeed?"). Include error states, validation failures, and retry loops. For each step, note whether it is a user action (click, type, swipe) or a system action (redirect, API call, email sent).
  3. Map Happy Path First, Then Edge Paths: Draw the ideal path from entry to success first. Then layer in alternative paths: what happens if the user is not logged in, if validation fails, if the session times out, if the payment is declined. Each branch should terminate in either a success state, an error recovery path, or an exit point.
  4. Annotate with Metrics and Risk Points: At each step, note the relevant metric: page view count, click-through rate, form completion rate, drop-off percentage. Flag high-friction steps where users are likely to abandon (multi-field forms, account creation walls, payment pages). Suggest specific optimizations for each risk point.
  5. Generate the Mermaid Diagram: Produce a clean Mermaid graph TD diagram using consistent node shapes: rounded rectangles () for pages/screens, diamonds {} for decisions, rectangles [] for actions, and double-circles (()) for start/end states. Use descriptive labels and color annotations via style directives to highlight error paths in red and success paths in green.
  6. Review and Iterate: Walk through the diagram with the user to verify completeness. Check that every branch terminates, that no orphan nodes exist, and that the flow accurately reflects the actual product behavior including edge cases.

Usage

Provide the agent with the user goal, the product context (web app, mobile app, or cross-platform), and any known screens or steps. Optionally include analytics data (drop-off rates, conversion percentages) for the agent to annotate onto the diagram. The agent will produce a Mermaid diagram with annotations.

Examples

Example 1: E-Commerce Checkout Flow

User Request: "Map the complete checkout flow for our e-commerce site. Users can be logged in or guests. We support credit card and PayPal."

Agent Response:

graph TD
    START(("Start: Cart Page")) --> REVIEW["Review Cart Items"]
    REVIEW --> EMPTY{Cart empty?}
    EMPTY -- Yes --> BROWSE["Redirect to Shop Page"]
    EMPTY -- No --> AUTH{User logged in?}

    AUTH -- Yes --> SHIPPING["Enter Shipping Address"]
    AUTH -- No --> CHOICE{Guest or Sign In?}
    CHOICE -- Sign In --> LOGIN["Login Page"]
    LOGIN --> LOGIN_OK{Login successful?}
    LOGIN_OK -- Yes --> SHIPPING
    LOGIN_OK -- No --> LOGIN_ERR["Show Error, Retry"]
    LOGIN_ERR --> LOGIN
    CHOICE -- Guest --> GUEST_EMAIL["Enter Email Address"]
    GUEST_EMAIL --> SHIPPING

    SHIPPING --> VALIDATE_ADDR{Address valid?}
    VALIDATE_ADDR -- No --> ADDR_ERR["Show Validation Errors"]
    ADDR_ERR --> SHIPPING
    VALIDATE_ADDR -- Yes --> DELIVERY["Select Delivery Method"]

    DELIVERY --> PAYMENT{Payment method?}
    PAYMENT -- Credit Card --> CC["Enter Card Details"]
    CC --> CC_VALID{Card valid?}
    CC_VALID -- No --> CC_ERR["Show Card Error"]
    CC_ERR --> CC
    CC_VALID -- Yes --> PROCESS

    PAYMENT -- PayPal --> PP["Redirect to PayPal"]
    PP --> PP_OK{PayPal authorized?}
    PP_OK -- No --> PP_ERR["PayPal Cancelled"]
    PP_ERR --> PAYMENT
    PP_OK -- Yes --> PROCESS

    PROCESS["Process Payment"] --> SUCCESS{Payment successful?}
    SUCCESS -- Yes --> CONFIRM["Order Confirmation Page"]
    CONFIRM --> END(("End: Order Complete"))
    SUCCESS -- No --> DECLINE["Show Decline Message"]
    DECLINE --> PAYMENT

    style BROWSE fill:#fef3c7,stroke:#f59e0b
    style LOGIN_ERR fill:#fee2e2,stroke:#ef4444
    style ADDR_ERR fill:#fee2e2,stroke:#ef4444
    style CC_ERR fill:#fee2e2,stroke:#ef4444
    style PP_ERR fill:#fee2e2,stroke:#ef4444
    style DECLINE fill:#fee2e2,stroke:#ef4444
    style CONFIRM fill:#d1fae5,stroke:#10b981
    style END fill:#d1fae5,stroke:#10b981

Metric annotations:

  • Cart → Shipping: ~72% proceed (28% abandon after viewing cart)
  • Shipping → Payment: ~85% proceed (address validation is low friction)
  • Payment → Confirmation: ~64% proceed (highest drop-off; consider saved payment methods)
  • Guest vs Sign In split: ~55% choose guest checkout (reduce friction by keeping guest option prominent)

Example 2: SaaS Onboarding Flow with Decision Points

User Request: "Map the onboarding flow for our project management tool. New users sign up, verify email, choose a plan, create their first project, and invite team members."

Agent Response:

graph TD
    START(("Entry: Landing Page")) --> SIGNUP["Sign Up Form"]
    SIGNUP --> VALIDATE{Form valid?}
    VALIDATE -- No --> FORM_ERR["Show Validation Errors"]
    FORM_ERR --> SIGNUP
    VALIDATE -- Yes --> EMAIL_SENT["Verification Email Sent"]

    EMAIL_SENT --> VERIFIED{Email verified?}
    VERIFIED -- No, 24h+ --> RESEND["Resend Verification"]
    RESEND --> VERIFIED
    VERIFIED -- Yes --> PLAN{Choose Plan}

    PLAN -- Free --> FREE["Free Plan Selected"]
    PLAN -- Pro --> BILLING["Enter Billing Info"]
    BILLING --> BILLING_OK{Payment valid?}
    BILLING_OK -- No --> BILL_ERR["Show Billing Error"]
    BILL_ERR --> BILLING
    BILLING_OK -- Yes --> PRO["Pro Plan Activated"]

    FREE --> PROFILE["Complete Profile"]
    PRO --> PROFILE

    PROFILE --> CREATE["Create First Project"]
    CREATE --> INVITE{Invite team members?}
    INVITE -- Yes --> TEAM["Enter Team Emails"]
    TEAM --> INVITES_SENT["Invitations Sent"]
    INVITES_SENT --> DASHBOARD
    INVITE -- Skip --> DASHBOARD["Dashboard - Onboarding Complete"]
    DASHBOARD --> END(("End: Active User"))

    style FORM_ERR fill:#fee2e2,stroke:#ef4444
    style BILL_ERR fill:#fee2e2,stroke:#ef4444
    style RESEND fill:#fef3c7,stroke:#f59e0b
    style DASHBOARD fill:#d1fae5,stroke:#10b981
    style END fill:#d1fae5,stroke:#10b981

Metric annotations and optimization notes:

  • Sign Up → Email Verified: ~68% verify within 1 hour. Send a reminder at 24 hours. Consider allowing limited access before verification to reduce early churn.
  • Plan Selection: ~80% choose Free initially. Offer a 14-day Pro trial without requiring billing info to increase Pro conversion.
  • Create First Project: Critical activation step. If the user does not create a project within 48 hours, trigger a guided tutorial email. Target: 60% activation within first session.
  • Invite Team: ~35% skip this step. Show the value of collaboration (e.g., "Teams complete projects 3x faster") to increase invite rates.

Best Practices

  • Always map the happy path before adding branches: Starting with the ideal path keeps the diagram readable. Layer in error states, edge cases, and alternative paths incrementally.
  • Use consistent node shapes: Reserve diamonds for decision points, rounded rectangles for screens/pages, and rectangles for user or system actions. This visual grammar makes flows scannable at a glance.
  • Terminate every branch: Every path in the diagram must end at a defined state — success, error recovery, or explicit exit. Orphan nodes indicate missing logic in the product.
  • Annotate with real data when available: Drop-off rates, conversion percentages, and session duration at each step transform a flow diagram from a planning artifact into an optimization tool.
  • Keep flows to one user goal per diagram: Combining "sign up," "purchase," and "manage account" in one diagram creates an unreadable mess. Map each goal separately, then link diagrams at shared entry/exit points.

Edge Cases

  • Circular flows (retry loops): Payment retries, form re-submissions, and re-authentication can create infinite loops in the diagram. Add a maximum retry count annotation (e.g., "Max 3 attempts, then redirect to support") and a terminal exit node for exhausted retries.
  • External system redirects (OAuth, PayPal, 3D Secure): When the user leaves the product for third-party authentication, mark the transition explicitly and account for three outcomes: success return, cancellation return, and timeout/no-return. Include a "user never returns" exit node.
  • A/B test variants: When the flow differs between test groups, create a decision node at the variant split labeled with the test name and variant identifiers, then map both paths to their respective outcomes.
  • Flows that span multiple sessions: Password reset (email → click link hours later) or email verification may not complete in one session. Mark the session boundary explicitly and indicate what triggers re-entry (email link, push notification, return visit).
  • Offline or degraded states in mobile apps: If the product works offline, map what happens when connectivity drops mid-flow: queued actions, error messages, and sync-on-reconnect behavior.

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

32.52%
按下载量换算487

Claude

30.32%
按下载量换算454

Cursor

17.82%
按下载量换算267

Gemini CLI

10.02%
按下载量换算150

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