Token导航 LogoToken导航

Agent Skill · 研究检索

reflex-dev(反射开发)

reflex-dev 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。

reflex-dev 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,权限为需要联网;许可证为 MIT,第三方扫描有未通过项,按安装量在研究检索类的 73,675 个 Skill 里排第 35,670。

总安装 2,046周安装 87GitHub Stars 2下载 717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
MIT

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:reflex-dev(反射开发)
来源仓库:https://github.com/silvainfm/claude-skills
仓库路径:skills/reflex-dev
安装命令:
npx skills add https://github.com/silvainfm/claude-skills --skill reflex-dev
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/silvainfm/claude-skills --skill reflex-dev

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。

reflex-dev 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。适用于研究检索类任务,如信息搜集、资料筛选和知识整理,尤其适合多源数据聚合与初步分析。

通过关键词、任务描述或来源仓库提供查询条件,返回结构化候选结果列表供进一步处理。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。可结合来源仓库、安装命令和原始 README 继续核验具体用法。

来源仓库内路径:skills/reflex-dev

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/silvainfm/claude-skills --skill reflex-dev;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

    它需要需要联网,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

reflex-dev 怎么安装?

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

reflex-dev 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。

reflex-dev 需要什么权限?

权限级别是需要联网。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。

reflex-dev 是免费的吗?

来源标注的许可证是 MIT。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

reflex-dev 用的人多吗?

公开数据是累计安装 2,046 次,近一周安装 87 次,下载 717 次,GitHub 2 星。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。

reflex-dev 安全吗?

第三方扫描有未通过项,风险较高。除非你确认过源码,否则不建议直接安装到常用环境。

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。
  • 3 个目录同时收录,说明这个 Skill 流传比较广。
  • 许可证明确为 MIT,商用和分发边界好判断。

先想清楚的地方

  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 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

  • 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
reflex-dev

反射开发 · 本页

2,0462需要联网MIT未通过

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

Reflex Development

Overview

Reflex is a full-stack Python framework for building web applications without writing JavaScript. Apps compile to a React frontend and FastAPI backend, with state management and event handlers running entirely in Python.

Architecture:

  • Frontend: Compiled to React (JavaScript) for UI rendering
  • Backend: FastAPI server running Python event handlers
  • Communication: WebSockets for real-time state updates
  • State: Server-side Python state synchronized to frontend

Core Concepts

State Management

State is a Python class that holds all mutable data and event handlers. All state variables must be JSON-serializable.

import reflex as rx

class AppState(rx.State):
    # State variables (any JSON-serializable type)
    count: int = 0
    items: list[str] = []
    user_name: str = ""

    # Event handlers - the ONLY way to modify state
    def increment(self):
        self.count += 1

    def add_item(self, item: str):
        self.items.append(item)

    # Computed vars (derived state)
    @rx.var
    def item_count(self) -> int:
        return len(self.items)

Key Rules:

  • State vars MUST be JSON-serializable (int, str, list, dict, bool, float)
  • Only event handlers can modify state
  • Use @rx.var decorator for computed/derived values
  • State is per-user session (isolated between users)

Components

Components are UI building blocks. Reflex provides 60+ built-in components.

import reflex as rx

def header() -> rx.Component:
    return rx.heading("My App", size="lg")

def counter_component(state: AppState) -> rx.Component:
    return rx.vstack(
        rx.text(f"Count: {state.count}"),
        rx.button("Increment", on_click=state.increment),
        spacing="4"
    )

Common Components:

  • Layout: rx.vstack, rx.hstack, rx.box, rx.container
  • Text: rx.heading, rx.text, rx.code
  • Input: rx.input, rx.text_area, rx.select, rx.checkbox
  • Interactive: rx.button, rx.link, rx.icon_button
  • Data: rx.table, rx.data_table, rx.list
  • Charts: rx.recharts.line_chart, rx.recharts.bar_chart, etc.

Event Handlers

Event handlers respond to user interactions and are the ONLY way to modify state.

class FormState(rx.State):
    form_data: dict[str, str] = {}

    # Simple event handler
    def handle_submit(self):
        print(f"Submitted: {self.form_data}")

    # Event handler with argument
    def update_field(self, field: str, value: str):
        self.form_data[field] = value

    # Async event handler (for API calls, DB queries)
    async def fetch_data(self):
        # Can use any Python library
        import httpx
        async with httpx.AsyncClient() as client:
            response = await client.get("https://api.example.com/data")
            self.data = response.json()

Event Triggers (connect components to handlers):

  • on_click: Button clicks
  • on_change: Input field changes
  • on_submit: Form submissions
  • on_mount: Component first renders
  • on_blur, on_focus: Input focus events

Project Structure

Standard Reflex app structure:

my_app/
├── my_app/
│   ├── __init__.py         # Empty
│   └── my_app.py           # Main app file (State + pages)
├── assets/                 # Static files (images, fonts, etc.)
├── .web/                   # Auto-generated frontend (don't edit)
├── rxconfig.py             # Reflex configuration
└── requirements.txt        # Python dependencies

Main App File Pattern

import reflex as rx

# 1. Define State
class State(rx.State):
    count: int = 0

    def increment(self):
        self.count += 1

# 2. Define Pages
def index() -> rx.Component:
    return rx.container(
        rx.heading("Welcome"),
        rx.button("Click", on_click=State.increment),
        rx.text(f"Count: {State.count}")
    )

def about() -> rx.Component:
    return rx.container(
        rx.heading("About"),
        rx.link("Home", href="/")
    )

# 3. Create App and Add Routes
app = rx.App()
app.add_page(index, route="/")
app.add_page(about, route="/about")

Common Patterns

Form Handling

class FormState(rx.State):
    name: str = ""
    email: str = ""

    def handle_submit(self, form_data: dict):
        self.name = form_data.get("name", "")
        self.email = form_data.get("email", "")

def form_page():
    return rx.form(
        rx.vstack(
            rx.input(name="name", placeholder="Name"),
            rx.input(name="email", placeholder="Email"),
            rx.button("Submit", type="submit"),
        ),
        on_submit=FormState.handle_submit,
    )

Data Tables

class DataState(rx.State):
    data: list[dict] = [
        {"id": 1, "name": "Alice", "age": 25},
        {"id": 2, "name": "Bob", "age": 30},
    ]

def data_table_page():
    return rx.data_table(
        data=DataState.data,
        columns=["id", "name", "age"],
        sort=True,
        search=True,
        pagination=True,
    )

File Upload

class UploadState(rx.State):
    async def handle_upload(self, files: list[rx.UploadFile]):
        for file in files:
            upload_data = await file.read()
            # Process file data
            outfile = f"./uploads/{file.filename}"
            with open(outfile, "wb") as f:
                f.write(upload_data)

def upload_page():
    return rx.vstack(
        rx.upload(
            rx.button("Select Files"),
            id="upload1",
        ),
        rx.button(
            "Upload",
            on_click=UploadState.handle_upload(rx.upload_files(upload_id="upload1"))
        ),
    )

Database Integration (with DuckDB)

import duckdb
import polars as pl

class DBState(rx.State):
    records: list[dict] = []

    async def load_data(self):
        # Use existing database connection
        conn = duckdb.connect("data/mydb.duckdb")
        df = conn.execute("SELECT * FROM mytable").pl()
        self.records = df.to_dicts()
        conn.close()

    async def insert_record(self, data: dict):
        conn = duckdb.connect("data/mydb.duckdb")
        conn.execute(
            "INSERT INTO mytable (name, value) VALUES (?, ?)",
            [data["name"], data["value"]]
        )
        conn.close()
        await self.load_data()  # Refresh

Styling & Layout

Inline Styling

rx.box(
    rx.text("Styled text"),
    bg="#1a5f9e",
    color="white",
    padding="4",
    border_radius="md",
)

Responsive Layout

rx.container(
    rx.responsive_grid(
        rx.box("Item 1", bg="blue"),
        rx.box("Item 2", bg="green"),
        rx.box("Item 3", bg="red"),
        columns=[1, 2, 3],  # 1 col mobile, 2 tablet, 3 desktop
        spacing="4",
    ),
    max_width="1200px",
)

Common Style Props

  • Layout: width, height, padding, margin, display
  • Colors: bg (background), color (text)
  • Typography: font_size, font_weight, text_align
  • Borders: border, border_radius, border_color
  • Spacing: spacing (for stacks), gap

Routing

Multiple Pages

app = rx.App()

# Route with parameters
@rx.page(route="/user/[id]")
def user_page() -> rx.Component:
    return rx.text(f"User ID: {State.router.page.params.get('id')}")

# Simple routes
app.add_page(index, route="/")
app.add_page(about, route="/about")

Navigation

# Links
rx.link("Go to About", href="/about")

# Programmatic navigation
def go_home(self):
    return rx.redirect("/")

Development Workflow

Initialize New App

pip install reflex
reflex init

Run Development Server

reflex run

App runs on http://localhost:3000 with auto-reload.

Common Commands

reflex run          # Start dev server
reflex export       # Build production bundle
reflex db init      # Initialize database (if using Reflex DB)
reflex db migrate   # Run migrations

Best Practices

  1. State Organization: Split large states into substates class AuthState(rx.State): user: str = "" class DataState(rx.State): items: list = []
  2. Component Reusability: Create reusable component functions def card(title: str, content: str) -> rx.Component: return rx.box(rx.heading(title, size="md"), rx.text(content), padding="4", border="1px solid #ddd",)
  3. Event Handler Performance: Use async for I/O operations async def fetch_data(self): # Async I/O won't block other users self.data = await some_api_call()
  4. Type Hints: Always type-hint state vars and event handlers count: int = 0 items: list[str] = [] def update_count(self, value: int) -> None: self.count = value

References

Documentation

Example Apps

See examples/ directory for complete working examples:

  • Simple counter app
  • Data table with CRUD operations
  • Form with validation
  • File upload and processing

Common Patterns Reference

See references/patterns.md for detailed examples of:

  • Authentication flows
  • Real-time updates
  • Complex form validation
  • Multi-step workflows
  • Data visualization with charts

平台分布

Antigravity

25.92%
按下载量换算186

Claude Code

23.55%
按下载量换算169

OpenCode

17.59%
按下载量换算126

Gemini CLI

11.73%
按下载量换算84

windsurf

6.62%
按下载量换算47

trae

2.95%
按下载量换算21

安全审计

Gen Agent Trust Hub

未通过

Socket

通过

Snyk

可疑

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。来源安全扫描存在提醒或未通过项,安装前请自行复核。

继续浏览同类 Skills