Token导航 LogoToken导航

Agent Skill · 研究检索

ralph-verify-acceptance-criteria(拉尔夫验证验收标准)

ralph-verify-acceptance-criteria 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。

ralph-verify-acceptance-criteria 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,权限为接触敏感数据;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 73,675 个 Skill 里排第 54,235。

总安装 514周安装 21GitHub Stars 公开资料未说明下载 166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:ralph-verify-acceptance-criteria(拉尔夫验证验收标准)
来源仓库:https://github.com/ralphcrisostomo/nuxt-development-skills
仓库路径:skills/ralph-verify-acceptance-criteria
安装命令:
npx skills add https://github.com/ralphcrisostomo/nuxt-development-skills --skill ralph-verify-acceptance-criteria
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/ralphcrisostomo/nuxt-development-skills --skill ralph-verify-acceptance-criteria

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。

用于查找、检索与验收标准相关的文档或规范要求。适合在项目交付前核对功能实现是否符合预期条件。可结合 PRD 或测试用例快速匹配关键指标。需确认数据来源可靠性,避免引用过时或错误信息。建议与实际代码变更交叉验证,确保一致性。

ralph-verify-acceptance-criteria 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/ralph-verify-acceptance-criteria

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/ralphcrisostomo/nuxt-development-skills --skil…;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

    它需要接触敏感数据,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

ralph-verify-acceptance-criteria 怎么安装?

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

ralph-verify-acceptance-criteria 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。

ralph-verify-acceptance-criteria 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

ralph-verify-acceptance-criteria 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。

ralph-verify-acceptance-criteria 用的人多吗?

公开数据是累计安装 514 次,近一周安装 21 次,下载 166 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。

ralph-verify-acceptance-criteria 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。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 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。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 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
ralph-verify-acceptance-criteria

拉尔夫验证验收标准 · 本页

514—接触敏感数据未明确通过

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

Ralph Verify Acceptance Criteria Skill

Verify that a user story implementation satisfies every acceptance criterion in scripts/ralph/prd.json. Produces a structured pass/fail report and optionally updates the PRD.

When to Use

  • After Ralph completes a user story — verify before marking passes: true
  • When the user asks to audit or re-verify existing implementations
  • To check if a specific acceptance criterion is met
  • To generate a verification report across all stories

Workflow

1. Load the PRD

Read scripts/ralph/prd.json from the project root. Parse the userStories array.

If the user specifies a story ID (e.g., "verify US-005"), target that story. If no story is specified, find the highest-priority story where passes: true that hasn't been verified yet, or ask the user which story to verify.

2. Detect UI Story

Before classifying individual criteria, determine if this story involves UI. Scan the story title, description, and acceptance criteria for any of these signals:

  • UI keywords: page, component, layout, modal, form, card, table, badge, button, tab, drawer, sidebar, header, nav, dialog, toast, skeleton, avatar, dropdown, menu, list, grid
  • File paths: .vue files, app/pages/, app/components/, app/layouts/
  • Nuxt UI components: UCard, UButton, UBadge, UTable, USkeleton, UModal, UInput, UForm, UDropdown, UTabs, UAvatar, etc.
  • Behavioral language: renders, displays, shows, visible, clicks, navigates, responsive, mobile, dark mode, screenshot

If any signal matches, mark the story as a UI story. Browser verification is mandatory for UI stories — it cannot be skipped or deferred.

3. Classify Each Acceptance Criterion

For each acceptance criterion, determine the verification strategy:

Pattern in criterionStrategyTool
"Create file", "Add file"File existsGlob/Read
"middleware", "composable", "component", "page", "layout"File exists + content checkRead + Grep
"Typecheck passes"Typecheckbun run lint via Bash
"tf:plan", "terraform", "Terraform module"Terraform planbun run tf:plan:staging
"GraphQL schema", "Add query", "Add mutation", "Add type"Schema checkGrep on schema.graphql
"resolver", "function module", "pipeline resolver"Terraform config checkGrep on main.tf and terraform/functions/
"Wire up", "datasource"Terraform wiringGrep on main.tf
"UCard", "UButton", "UBadge", "USkeleton" etc.Component usageGrep in target Vue file
General behavior descriptionsCode reviewRead the relevant file and verify logic

Browser verification is not a per-criterion classification — it is driven by the UI story detection in Step 2.

4. Execute Verifications

Run checks in this order. Browser verification runs first because it catches the most impactful issues (broken renders, missing elements, runtime errors) that static checks miss.

A. Browser Verification (UI Stories)

Run this phase first for any story detected as UI in Step 2. Use the /dev-browser skill:

  1. Start the dev server if not running: bun run dev &
  2. Start the browser: ./skills/dev-browser/server.sh &
  3. Auth setup:

- Regular pages: set auth-role=user cookie + ?dev-bypass query param - Admin pages: use demo account (.env → DEV_DEMO_EMAIL / DEV_DEMO_PASSWORD)

  1. Navigate to the target page
  2. Verify checklist:

- Page renders without console errors - Expected UI elements are visible (cards, tables, buttons, badges, forms) - Key interactions work (clicks, navigation, form submission) - Responsive layout if criteria mention mobile

  1. Take screenshots as evidence
B. File Existence Checks

For criteria mentioning file creation, use Glob to confirm files exist. Read key files to verify they contain expected patterns (component names, function signatures, imports).

C. Code Content Checks

For behavioral criteria, Read the relevant source files and verify:

  • Expected functions/methods exist
  • Required imports or composable calls are present
  • Correct patterns are used (e.g., definePageMeta({layout: 'admin', middleware: ['admin']}))
  • GraphQL query strings are defined
  • Resolver functions contain expected DynamoDB operations
D. GraphQL Schema Checks

For schema-related criteria, Grep terraform/envs/staging/schema.graphql for:

  • Type definitions (type AdminStats)
  • Query/mutation declarations (getAdminStats, adminDeleteUser)
  • Field definitions with correct types
  • Input arguments
E. Terraform Infrastructure Checks

For infrastructure criteria:

  • Grep terraform/envs/staging/main.tf for module declarations
  • Grep terraform/envs/staging/lambda_function.tf for Lambda resources
  • Check terraform/functions/ for resolver JS files
  • Check terraform/lambda/src/ for Lambda handler code
  • If the user wants a full terraform validation, run bun run tf:plan:staging (requires AWS credentials)
F. TypeScript / Lint Check

For "Typecheck passes" criteria, run:

bun run lint 2>&1 | tail -20

Check exit code. If it fails, capture and report the errors.

5. Generate Verification Report

Output a structured report. Browser verification appears first to reflect execution priority:

## Verification Report: US-XXX — [Story Title]

### Acceptance Criteria Results

| # | Criterion | Status | Evidence |
|---|-----------|--------|----------|
| 1 | Browser verification | PASS | Screenshot: page renders, elements visible |
| 2 | Create middleware file | PASS | File exists at `app/middleware/admin.ts` |
| 3 | Non-admin redirect | PASS | Line 8: `navigateTo('/')` with toast |
| 4 | Typecheck passes | PASS | `bun run lint` exit code 0 |

### Summary
- **Passed:** 4/4
- **Failed:** 0/4
- **Overall:** PASS

For failed criteria, include specific details about what's missing or incorrect, with file paths and line numbers.

6. Update PRD (Optional)

If all criteria pass and the user confirms, update scripts/ralph/prd.json: set "passes": true and add "verifiedAt": "<ISO timestamp>". If any criteria fail, do NOT update — list remediation steps instead.

Batch Verification Mode

When asked to "verify all stories" or "audit the PRD":

  1. Load all stories from prd.json
  2. Filter to stories where passes: true (already claimed complete)
  3. Run verification on each story sequentially
  4. Output a summary table:
| Story | Title | Claimed | Verified | Issues |
|-------|-------|---------|----------|--------|
| US-001 | Admin middleware | PASS | PASS | - |
| US-003 | Admin layout | PASS | FAIL | Missing mobile drawer test |

Quick Check Mode

When the user says "quick verify" or "verify without browser", static checks (file existence, code patterns, schema, terraform, typecheck) run immediately. For UI stories, browser verification is marked as DEFERRED in the report — the story cannot be marked passes: true until browser verification completes. Run deferred browser checks later with "verify browser US-XXX".

Key Project Paths

AreaPath
Pagesapp/pages/
Componentsapp/components/
Middlewareapp/middleware/
Composablesapp/composables/
Layoutsapp/layouts/
GraphQL queriesapp/graphql/
Schematerraform/envs/staging/schema.graphql
Resolver functionsterraform/functions/
Lambda handlersterraform/lambda/src/
Terraform configterraform/envs/staging/main.tf
Lambda TF configterraform/envs/staging/lambda_function.tf
Types / PRDtypes/, scripts/ralph/prd.json

平台分布

Codex

33.26%
按下载量换算55

Claude

33.25%
按下载量换算55

Cursor

19.6%
按下载量换算33

Gemini CLI

10.41%
按下载量换算17

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

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