Token导航 LogoToken导航

Agent Skill · 研究检索

eventmodeling-identifying-inputs(事件建模识别输入)

eventmodeling-identifying-inputs 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。

eventmodeling-identifying-inputs 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 73,674 个 Skill 里排第 56,981。

总安装 442周安装 19GitHub Stars 公开资料未说明下载 155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:eventmodeling-identifying-inputs(事件建模识别输入)
来源仓库:https://github.com/trogonstack/agentskills
仓库路径:skills/eventmodeling-identifying-inputs
安装命令:
npx skills add https://github.com/trogonstack/agentskills --skill eventmodeling-identifying-inputs
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/trogonstack/agentskills --skill eventmodeling-identifying-inputs

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。

eventmodeling-identifying-inputs 区分用户输入命令与系统内部触发动作。防止错误归因导致后续追踪链条断裂。eventmodeling-identifying-inputs 属于研究检索类 Skill,可作为该场景下的辅助能力补充。建立角色目录,关联输入源与处理逻辑,增强可观测性。

适用于复杂自动化流程较多的业务系统建模。需结合 UI 交互原型或现有 API 文档进行交叉验证。

来源仓库内路径:skills/eventmodeling-identifying-inputs

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/trogonstack/agentskills --skill eventmodeling-…;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

eventmodeling-identifying-inputs 怎么安装?

在终端执行 npx skills add https://github.com/trogonstack/agentskills --skill eventmodeling-identifyin…。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

eventmodeling-identifying-inputs 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。

eventmodeling-identifying-inputs 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。

eventmodeling-identifying-inputs 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。

eventmodeling-identifying-inputs 用的人多吗?

公开数据是累计安装 442 次,近一周安装 19 次,下载 155 次。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。

eventmodeling-identifying-inputs 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
eventmodeling-identifying-inputs

事件建模识别输入 · 本页

442—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

Identifying Inputs

Interview Phase (Optional)

When to Interview: Skip if the user has already identified UI actions/commands and processor triggers. Interview when it's unclear which actions are user-initiated vs. processor-automated.

Interview Strategy: Separate UI-driven commands from processor-driven commands before cataloging inputs. Mixing them leads to incorrect role attribution, which breaks the Role Catalog traceability that downstream steps depend on.

Critical Questions

  1. Automation Level (Impact: Determines which commands are UI-triggered vs. processor-triggered)

- Question: "Are there actions that should be: (A) User-initiated only, (B) Processor-automated, (C) Mix of both?" - Why it matters: Knowing automation vs. manual separates command types - Follow-up triggers: If (C) → ask "Which specific user actions trigger automation? What does the processor decide on its own?"

  1. External System Triggers (Impact: Determines if there are processor commands from webhooks/integrations)

- Question: "Will commands be triggered by: (A) UI only, (B) External webhooks (payments, notifications, etc.), (C) Scheduled processors, (D) All of above?" - Why it matters: External triggers are processor commands, not UI commands - Follow-up triggers: If (B) or (D) → ask which external systems send webhooks and what data they include

Interview Flow

Conditional Entry:

If user has provided:
  - UI actions already listed per storyboard screen
  - AND processor triggers identified with source systems named
  - AND it's clear which role/actor initiates each action

Then: Skip interview, proceed directly to command identification

Else: Conduct interview

Phase 1: Trigger Classification (Question 1)

  • Establish which commands come from human actors vs. automated processors
  • Confirm Role Catalog from Step 1 is available for attribution

Phase 2: External Triggers (Question 2)

  • Identify all external system integrations that issue commands
  • Confirm whether scheduled jobs or event-driven processors exist

Capturing Interview Findings

Append findings to the project's event modeling file:

File: .trogonai/interviews/[project-name]/EVENTMODELING.md

Use Write tool to add/update this section:

## 4. Identifying Inputs (eventmodeling-identifying-inputs)

### Automation Classification
[From Q1: Which actions are user-initiated vs. processor-automated?]

### External System Triggers
[From Q2: Which external systems trigger commands? Webhook formats?]

### Command Attribution Summary
- UI-issued commands: [list with role from Role Catalog]
- Processor-issued commands: [list with source system]

Update Interview Trail:

| 4 | eventmodeling-identifying-inputs | Done | UI commands, processor commands, role attribution |

Workflow

Given UI storyboards and event timeline, identify all inputs.

PREREQUISITE: The Role Catalog from Step 1 (eventmodeling-brainstorming-events) must exist. Every command identified below MUST be attributed to a specific role or system actor from that catalog.

1. Extract Commands from UI Actions

For each user action in storyboard, create a command attributed to a specific role:

Storyboard: Order Creation Screen
User action: Click "Create Order" button
  ↓
Command: CreateOrder
Input data from form:
    - customerId
    - items[] (product selections + quantities)
    - shippingAddress
Validation:
    - customerId must exist
    - items must not be empty
    - quantities must be > 0
Produces event: OrderCreated

2. Identify Processor Triggers

Identify automation-triggered commands:

Processor trigger: Payment gateway webhook received
  ↓
Command: AuthorizePayment (from Processor, not UI)
Input data from webhook:
    - orderId
    - paymentId
    - authorizationCode
Validation:
    - orderId must exist and be in Confirmed state
    - authorizationCode must be valid
Produces event: PaymentAuthorized

2b. Understand the Processor "Todo List" Pattern

Processors don't directly process events—they maintain a todo list driven by events:

Event Stream (Domain events):
PaymentAuthorized → triggers Inventory system

Processor: InventoryReserver

Todo List:
When PaymentAuthorized event arrives:
    1. Add item to todo: "Reserve inventory for order-123"

Processor Logic (continuously):
FOR EACH todo item IN todo_list:
    - Check if inventory available
    - If yes: Reserve inventory, produce InventoryReserved event, mark done
    - If no: Produce InventoryFailed event, mark failed
    - If error: Keep in todo for retry

Example:
Event: PaymentAuthorized(orderId=order-123, items=[{prodId: P1, qty: 2}])
    ↓
Todo added: Reserve P1 qty 2
    ↓
Processor checks: P1 has 5 available, need 2
    ↓
Action: Reserve 2 units
    ↓
Event produced: InventoryReserved(orderId=order-123, reserved=[...])
    ↓
Todo marked done

Key insight: Processors are reactive. They listen for events and create todo items, then execute those todos by issuing commands that produce new events.

2c. Document Processor Automation (Gears Symbol)

Show which commands come from automation vs. user actions:

Command Catalog with Role Attribution (from Role Catalog):

UI-Issued Commands (attributed to specific human roles):
  1. CreateOrder (Order Entry screen) [ Customer]
  2. ConfirmOrder (Confirmation screen) [ Customer]
  3. CancelOrder (Status screen) [ Customer]
  4. RequestReturn (Order page) [ Customer]
  5. OverrideOrderStatus (Admin panel) [ Support Agent]

Processor-Issued Commands (attributed to system actors):
  6. AuthorizePayment (Payment gateway webhook) [ Payment Gateway]
  7. ReserveInventory (Triggered by PaymentAuthorized) [ Inventory System]
  8. CreateShipment (Triggered by InventoryReserved) [ Fulfillment System]
  9. NotifyCustomer (Triggered by multiple events) [ Notification Service]

Validation: Every command MUST have a role/actor attribution. If a command says [User] instead of a specific role name, it's incomplete — go back to the Role Catalog and assign the correct role.

3. Document Command Specifics

For each command, define structure:

Command: ConfirmOrder
Source: UI (user clicks button)
Input:
    orderId: string (from URL/context)
    paymentMethod: enum ('card' | 'transfer')
    [paymentDetails]: depends on method

Validation rules:
    - Order must exist
    - Order must be in Draft state
    - Payment method must be supported
    - Funds must be available (pre-check)

Preconditions (from stream state):
    - OrderCreated event exists
    - No ConfirmOrder previously processed

Success result: OrderConfirmed event

Failure results:
    - "Order not found" → Command rejected, no event
    - "Order already confirmed" → Command rejected, no event
    - "Payment method not supported" → Command rejected, no event

4. Create Command Catalog

List all commands the system accepts:

Command Catalog: Order System

### UI-Issued Commands

1. CreateOrder
   Source: User (Order Entry screen)
   Input: customerId, items[], shippingAddress
   Produces: OrderCreated event

2. ConfirmOrder
   Source: User (Confirmation screen)
   Input: orderId, paymentMethod
   Produces: OrderConfirmed event

3. CancelOrder
   Source: User (Status screen)
   Input: orderId, reason
   Produces: OrderCancelled event

### Processor-Issued Commands

4. AuthorizePayment
   Source: Payment Processor (webhook)
   Input: orderId, paymentId, authCode
   Produces: PaymentAuthorized event

5. FailPayment
   Source: Payment Processor (webhook)
   Input: orderId, paymentId, reason
   Produces: PaymentFailed event

6. ReserveInventory
   Source: Inventory Processor (triggered by PaymentAuthorized)
   Input: orderId, items[]
   Produces: InventoryReserved event

7. CreateShipment
   Source: Fulfillment Processor (triggered by InventoryReserved)
   Input: orderId, items[]
   Produces: OrderShipped event

5. Map Data Sources

Document where each command input comes from:

Command: ConfirmOrder

Data origin matrix:
  orderId
    ↑ Source: UI context (from OrderCreated, displayed to user)
    ↑ Captured: Hidden in URL or session
    ↑ Validation: Must match Order from stream

  paymentMethod
    ↑ Source: UI form selection
    ↑ Captured: User selects checkbox/radio
    ↑ Validation: Must be in allowed list

[paymentDetails] (conditional)
    ↑ Source: Depends on paymentMethod
    ↑ For 'card': Card number, CVV, expiry (from payment form)
    ↑ For 'transfer': Bank account, routing number (from form)
    ↑ Validation: Format and validity checks

6. Identify Implicit Context

Document what comes from stream state:

Command: ShipOrder
Explicit input (from UI/Processor):
    orderId
    shipmentId (from fulfillment system)

Implicit context (from stream state):
    Order must exist
    Order must be in InventoryReserved state
    Payment must be authorized (from PaymentAuthorized event)
    Inventory must be reserved (from InventoryReserved event)

These implicit checks use stream state:
    currentState.orderId === orderId
    currentState.status === 'InventoryReserved'
    currentState.paymentId exists
    currentState.shipmentId can be set

Output Format

Present as:

# Inputs: [Domain Name]

## Commands Summary

| Command | Role/Actor | Source | Trigger | Input | Event |
|---------|------------|--------|---------|-------|-------|
| CreateOrder | Customer | UI | User action | customerId, items, address | OrderCreated |
| ConfirmOrder | Customer | UI | User action | orderId, paymentMethod | OrderConfirmed |
| CancelOrder | Customer | UI | User action | orderId, reason | OrderCancelled |
| AuthorizePayment | Payment Gateway | Processor | Webhook | orderId, paymentId | PaymentAuthorized |
| ReserveInventory | Inventory System | Processor | PaymentAuthorized event | orderId, items | InventoryReserved |
| ShipOrder | Fulfillment System | Processor | InventoryReserved event | orderId, shipmentId | OrderShipped |

---

## Detailed Commands

### Command: CreateOrder

**Source**: User (Order Entry screen)

**Input Data**:
- customerId: string
- items: Array<{productId: string, quantity: number}>
- shippingAddress: {street, city, state, zip}

**Validation**:
- customerId must exist in system
- items array must not be empty
- quantities must be > 0
- address fields must be non-empty

**Preconditions** (from stream state):
- Stream Order:X does not exist yet

**Success**: Produces OrderCreated event

**Failure**: Command rejected, no event
- "Customer not found"
- "Items invalid"
- "Address incomplete"

--- [Repeat for each command]

---

## Data Completeness Check

### Data Input → Event

Verify every command input becomes event data:

| Command Input | Event Data | Status |
|---------------|-----------|--------|
| customerId | orderId |  Stored in OrderCreated |
| items | items |  Stored in OrderCreated |
| shippingAddress | shippingAddress |  Stored in OrderCreated |

### Missing Data

Document any input that doesn't make it to events:
- None identified

---

## Processor Commands

Document all processor-triggered commands:
[List each with source system and trigger condition]

Quality Checklist

  • Every UI action maps to a command
  • Every processor action maps to a command
  • Every command is attributed to a specific role/actor from the Role Catalog
  • No command uses generic "User" — must name the specific role (Customer, Seller, Admin, etc.)
  • Every command input is documented
  • Every input validates against rules
  • Preconditions from stream state are explicit
  • Success and failure outcomes documented
  • Implicit context from stream state is identified
  • No undocumented commands exist
  • Command naming is consistent and clear
  • Processor triggers are explicit
  • Processor todo list pattern explained for each automation
  • Event-to-todo triggering mechanism documented
  • Automation marked with [AUTO] to distinguish from user actions [USER]
  • Processor failure/retry handling specified

Key Principles

  1. Source Clarity: Every command comes from UI or Processor
  2. Input Completeness: All needed data captured
  3. Validation Explicit: All rules documented
  4. State Awareness: Preconditions from stream state clear
  5. Event Mapping: Every input becomes event data

Common Patterns

User Command Pattern

User action on UI screen
  ↓
Captures form/selection data
  ↓
Validation checks
  ↓
Command issued
  ↓
Event created or rejection

Processor Command Pattern

External event/webhook received
  ↓
Triggers processor logic
  ↓
Processor validates and decides
  ↓
Command issued (if valid)
  ↓
Event created or decision recorded

Conditional Input Pattern

Command: PaymentConfirm
Input:
  - paymentMethod (user selected)
  - paymentDetails (conditional on method)
    If method='card': cardNumber, CVV, expiry
    If method='transfer': bankAccount, routingNumber

平台分布

Codex

36.45%
按下载量换算56

Claude

30.95%
按下载量换算48

Cursor

20.1%
按下载量换算31

Gemini CLI

10.14%
按下载量换算16

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