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

fbp-typesfbp 类型

Agent Skill

fbp-types 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

346

周安装

14

GitHub Stars

公开资料未说明

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

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:fbp-types(fbp 类型)
来源仓库:https://github.com/constructive-io/constructive-skills
仓库路径:skills/fbp-types
安装命令:
npx skills add https://github.com/constructive-io/constructive-skills --skill fbp-types
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/constructive-io/constructive-skills --skill fbp-types

简介

提供 Flow-Based Programming 的 TypeScript 类型定义与 GraphSchemata 规范。

  • 核心原则包括显式优于魔法、子图可组合与稳定内容寻址结构。
  • 所有实体均为节点,支持嵌套图结构与边连接关系建模。
  • 强调数据流与控制配置的明确分离,提升系统可维护性。
  • fbp-types 属于前端设计类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Flow-Based Programming TypeScript types and GraphSchemata specification.

Installation

pnpm add @fbp/types

Overview

@fbp/types provides TypeScript types for a Houdini-inspired, merkle-friendly graph specification. The core design principles are explicitness over magic, composable subgraphs, stable content-addressable structure, and clear separation between dataflow and configuration.

Core Concepts

Everything is a Node

The entire document is a single Node. Nodes may contain other nodes and edges, forming nested graphs (subnets). A node becomes a subnet simply by containing nodes[] and edges[]. There is no separate "graph" type.

Identity Model

Node identity equals name within parent scope. Names may be hierarchical paths by convention (some/subnet/node). Rename or move equals identity change. This model is compatible with merkle/content-addressable versioning.

Type Definitions

Graph

interface Graph {
  name?: string;
  nodes: Node[];
  edges: Edge[];
  definitions?: NodeDefinition[];
  meta?: NodeMeta;
}

Node

interface Node {
  name: string;               // Unique within parent scope
  type: string;               // References a NodeDefinition
  meta?: NodeMeta;            // Position and metadata
  props?: PropValue[];        // Property values
  nodes?: Node[];             // Child nodes (makes this a subnet)
  edges?: Edge[];             // Edges within this scope
}

Edge

interface Edge {
  src: PortRef;               // Source (output port)
  dst: PortRef;               // Destination (input port)
  channel?: string;           // Optional channel (default: "main")
}

interface PortRef {
  node: string;               // Node name within scope
  port: string;               // Port name
}

NodeDefinition

interface NodeDefinition {
  type: string;               // Unique identifier (e.g., "math/add")
  context?: string;           // Namespace (e.g., "math", "ui")
  category?: string;          // Palette category
  inputs?: PortDef[];         // Input port definitions
  outputs?: PortDef[];        // Output port definitions
  props?: PropDef[];          // Property definitions
  icon?: string;
  description?: string;
}

Port and Prop Definitions

interface PortDef {
  name: string;
  type?: string;              // Data type (e.g., "string", "number", "any")
  multi?: boolean;            // Accepts multiple connections
  description?: string;
}

interface PropDef {
  name: string;
  type?: string;
  default?: any;
  description?: string;
}

interface PropValue {
  name: string;
  value?: any;
  ref?: boolean;              // If true, value is a reference path
}

Ports (No Lanes)

Ports are named, typed, and singular. There is no lane or index dimension. Branching uses multiple named output ports (e.g., true, false, error). Fan-in is never implicit — multiple values must be combined via explicit nodes (Merge, Collect, etc.).

Props vs Dataflow

Props are not dataflow. They configure node behavior and are not connected by edges. They may reference other parameters via Ref:

{ "ref": "../config/apiKey" }

Boundary Nodes

Boundary nodes define a subnet's interface:

Node TypePurposeProperties
graphInputInput portportName, dataType
graphOutputOutput portportName, dataType
graphPropConfiguration propertypropName, dataType, default

Reserved names @in, @out, @props are used for canonical boundary nodes.

Channels

Edges may specify a channel for namespacing:

{
  src: { node: 'A', port: 'out' },
  dst: { node: 'B', port: 'in' },
  channel: 'error'
}

Default channel is "main". Non-main channels are for error routing, control edges, metadata propagation.

Validation Rules

  • Node names must be unique within a scope
  • Node names starting with @ are reserved for boundary nodes
  • Port names must be unique within inputs, outputs, and props
  • Edges must reference existing ports
  • Fan-in requires explicit merge nodes
  • Branching requires explicit output ports

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

37.41%
按下载量换算41

Claude

27.64%
按下载量换算30

Cursor

19.68%
按下载量换算21

Gemini CLI

10.17%
按下载量换算11

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