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clear审计提醒

recovery-social-features恢复社交功能

Agent Skill

用于辅助前端页面、组件、样式和交互逻辑的开发与维护。它适合让 Agent 生成或审查 React、Next.js、Vue、Tailwind、CSS 等相关代码,整理组件结构,或定位布局和性能问题。使用时需要结合项目现有设计系统、路由和构建方式,避免只生成孤立片段;涉及页面改动时,应配合本地预览和构建检查确认视觉效果。

总安装

2,012

周安装

83

GitHub Stars

98

下载量

657
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

3

许可证

MIT

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:recovery-social-features(恢复社交功能)
来源仓库:https://github.com/erichowens/some_claude_skills
仓库路径:skills/recovery-social-features
安装命令:
npx skills add https://github.com/erichowens/some_claude_skills --skill recovery-social-features
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/erichowens/some_claude_skills --skill recovery-social-features

简介

recovery-social-features 用于辅助前端页面、组件、样式和交互逻辑的开发与维护,适合在 Codex、Claude、Cursor、Gemini CLI 中生成或审查前端代码。

  • 适用于 React、Next.js、Vue、Tailwind、CSS 等相关技术栈的组件开发。
  • 可生成组件结构或定位布局和性能问题。
  • 安装命令:npx skills add https://github.com/erichowens/some_claude_skills --skill recovery-social-features。
  • 需结合项目现有设计系统和构建方式使用,避免生成孤立片段。

SKILL.md

Recovery-Focused Social Features

Build privacy-first social features for addiction recovery apps. These patterns prioritize anonymity, safety, and the unique relationship structures in recovery communities.

When to Use

USE this skill for:

  • Sponsor/sponsee relationship systems
  • Recovery-focused group features (meeting groups, accountability circles)
  • Privacy-first friend connections with mutual consent
  • Safe messaging between recovery peers
  • Anonymity-preserving profile systems

DO NOT use for:

  • General social media patterns → use standard social feature docs
  • Dating or romantic connection features → not appropriate for recovery context
  • Public-facing profiles → recovery apps should default to privacy
  • Content recommendation algorithms → use recovery-community-moderator for content safety

Core Principles

Privacy by Default

Recovery apps handle sensitive data. Default to maximum privacy, let users opt into visibility.

interface PrivacySettings {
  profileVisibility: 'private' | 'friends' | 'community';
  showSobrietyDate: boolean;
  showProgram: boolean;        // AA, NA, CMA, etc.
  showLocation: 'none' | 'city' | 'region';
  allowMessages: 'none' | 'friends' | 'sponsors' | 'all';
  anonymousInGroups: boolean;  // Use display name only
}

// Default to most private
const DEFAULT_PRIVACY: PrivacySettings = {
  profileVisibility: 'friends',
  showSobrietyDate: false,
  showProgram: false,
  showLocation: 'none',
  allowMessages: 'friends',
  anonymousInGroups: true,
};

Anonymity Support

Many users need complete anonymity. Support display names separate from real identity.

interface Profile {
  id: string;
  // Private - never exposed publicly
  email: string;
  realName?: string;

  // Public-facing identity
  displayName: string;        // "JohnD" or "GratefulMember"
  avatarType: 'initials' | 'icon' | 'custom';
  avatarIcon?: string;        // Predefined icon set

  // Recovery-specific
  sobrietyDate?: Date;
  programs: ('aa' | 'na' | 'cma' | 'smart' | 'refuge' | 'other')[];
  homeGroup?: string;         // Primary meeting
}

Feature Overview

Sponsor/Sponsee Relationships

The sponsor relationship is hierarchical and private. Only the two parties should know about it.

Key concepts:

  • Invite-based connection (sponsor generates code, sponsee accepts)
  • Time-limited invite codes (24h expiration)
  • Private by design - no public visibility
  • One sponsor per program, unlimited sponsees

Hooks provided:

  • useSponsorInvite() - Generate and accept invite codes
  • useSponsorRelationships() - List sponsors and sponsees

Components:

  • GenerateSponsorInvite - Sponsor creates shareable code
  • AcceptSponsorInvite - Sponsee enters code to connect
  • SponsorDashboard - View all sponsor/sponsee relationships
See: references/sponsor-sponsee.md for full implementation

Meeting-Based Groups

Groups that form organically around meetings. Ephemeral by default but can be made permanent.

Key concepts:

  • Tied to specific meetings or standalone
  • Ephemeral groups auto-delete after 24 hours
  • Visibility options: public, private, invite-only
  • Member limits prevent overcrowding

Group Settings:

interface GroupSettings {
  name: string;
  meetingId?: string;           // Link to meeting
  visibility: 'public' | 'private' | 'invite';
  ephemeral: boolean;           // Auto-delete after 24h
  maxMembers: number;           // Member limit
}
VisibilityWho Can SeeWho Can Join
publicAnyoneAnyone
privateMembers onlyInvite only
inviteMembers onlyHas invite code

Hooks provided:

  • useMeetingGroup() - Create, join, leave groups

Components:

  • QuickMeetingGroup - One-tap group creation at meetings
See: references/groups.md for full implementation

Friend Connections

Peer-to-peer connections without hierarchy. Mutual consent required.

Key concepts:

  • Request/accept flow (no auto-follows)
  • Real-time updates via Supabase subscriptions
  • Blocking supported (one-way, discreet)
  • Status: pending, accepted, blocked

Hooks provided:

  • useFriendships() - Full friendship management with real-time sync

Components:

  • FriendRequestButton - Context-aware add/pending/friends states
  • PendingFriendRequests - Accept/decline UI
See: references/friendships.md for full implementation

Safe Messaging

Recovery-appropriate messaging with crisis detection and safety features.

Key concepts:

  • Real-time message delivery via Supabase
  • Crisis keyword detection (non-blocking, shows resources)
  • Soft-delete (messages hidden, not destroyed)
  • Privacy-first (no read receipts by default)

Crisis Keywords (trigger resource prompt):

const CRISIS_KEYWORDS = [
  // Suicidal ideation
  'suicide', 'kill myself', 'want to die', 'end it all',
  // Relapse indicators
  'relapse', 'using again', 'fell off the wagon',
  // Self-harm
  'hurt myself', 'cutting', 'self-harm',
];

Best Practices:

  1. Non-blocking - Crisis prompts suggest resources, don't block messages
  2. Privacy-first - Don't log or report crisis keywords automatically
  3. Helpful tone - Gentle, non-judgmental language
  4. Direct resources - Link to crisis page, not external sites
  5. Offline capable - Cache crisis resources for offline access

Hooks provided:

  • useMessages() - Real-time message thread with Supabase subscriptions

Components:

  • MessageInput - Input with crisis detection overlay
See: references/messaging.md for full implementation

Accountability Features

Sharing recovery progress with trusted connections.

Check-In Sharing:

  • Share daily check-ins with selected sponsors
  • HALT tracking (Hungry, Angry, Lonely, Tired)
  • Mood and gratitude logging

Sobriety Visibility Settings:

LevelWho Can SeeUse Case
privateOnly selfMaximum privacy
sponsorsSelf + sponsorsAccountability focus
friendsSelf + sponsors + friendsPeer support
communityAll app usersPublic milestone celebrations

HALT Check-In Data:

interface DailyCheckIn {
  date: string;
  mood: 1 | 2 | 3 | 4 | 5;  // 1=worst, 5=best
  halt: {
    hungry: boolean;
    angry: boolean;
    lonely: boolean;
    tired: boolean;
  };
  gratitude?: string;
  notes?: string;
}

Components:

  • ShareCheckIn - Select sponsors to share with
  • SobrietyVisibility - Privacy level picker
See: references/accountability.md for full implementation

Safety & Moderation

Content moderation and user blocking for safe communities.

Moderation Categories:

CategoryDescriptionAction
crisisSuicidal ideation, self-harmShow resources, don't block
sourcingDrug seeking, dealingBlock + flag for review
harassmentPersonal attacks, threatsBlock + flag for review
spamPromotional contentBlock
explicitSexual/graphic contentBlock

Blocking Behavior:

  • Blocked user cannot send messages
  • Blocked user cannot see blocker's profile
  • Blocked user cannot see blocker in groups
  • Existing messages are hidden (not deleted)
  • Blocking is one-way (blocked user doesn't know)

RLS Policy Pattern:

-- Hide content from blocked users
CREATE POLICY "Hide messages from blocked users" ON messages
  FOR SELECT USING (
    NOT EXISTS (
      SELECT 1 FROM friendships
      WHERE status = 'blocked'
      AND (
        (requester_id = auth.uid() AND addressee_id = sender_id)
        OR (addressee_id = auth.uid() AND requester_id = sender_id)
      )
    )
  );

Hooks provided:

  • useContentModeration() - Check content against moderation API
  • useBlocking() - Block/unblock users, check block status
See: references/moderation.md for full implementation

Quick Reference

FeaturePrivacy DefaultWho Can See
ProfileFriends onlyConfigurable
Sobriety dateHiddenConfigurable
Sponsor relationshipPrivateOnly the two parties
Group membershipGroup membersConfigurable per group
MessagesParticipants onlyNever public
Check-insPrivateOpt-in sharing

Database Schema

See supabase-admin/references/social-schema.md for complete Supabase schema including:

  • Friendships table with RLS
  • Sponsorships with invite codes
  • Groups and group members
  • Conversations and messages
  • Real-time subscription patterns

References

Detailed implementations in /references/:

FileContents
sponsor-sponsee.mduseSponsorInvite hook, invite UI components, SponsorDashboard
groups.mduseMeetingGroup hook, QuickMeetingGroup component
friendships.mduseFriendships hook with real-time, friend request UI
messaging.mduseMessages hook, MessageInput with crisis detection
accountability.mdShareCheckIn, SobrietyVisibility components
moderation.mduseContentModeration, useBlocking hooks, RLS policies

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

04

需要参考平台分布和安装热度时

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

补充不同宿主或平台的使用分布数据

能力 5

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

Claude Code

29.51%
按下载量换算194

windsurf

26.42%
按下载量换算174

Antigravity

18.22%
按下载量换算120

OpenCode

12.52%
按下载量换算82

Gemini CLI

8.06%
按下载量换算53

Codex

4.03%
按下载量换算26

安全审计

Gen Agent Trust Hub

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 warning/failed 结果,不能写成本站确认安全。

来源信息

继续浏览同类 Skills