编程

Decision Fatigue Reliever

试用

Helps you make quick decisions on daily trivial matters (food, clothing, activities, shopping), reducing choice anxiety and freeing up cognitive resources. U...

它能做什么

Helps you make quick decisions on daily trivial matters, reducing choice anxiety and freeing up cognitive resources.

技能文档

Decision Fatigue Reliever

Overview

Helps you make quick decisions on daily trivial matters, reducing choice anxiety and freeing up cognitive resources.

"Let AI handle the trivial decisions, save your energy for what truly matters"


Triggers

Natural Language Triggers

Decision TypeExample Triggers
🍽️ What to Eat"What should I eat for lunch", "Help me decide dinner", "Suggest breakfast", "Don't know what to eat"
👔 What to Wear"What should I wear today", "Help me pick an outfit", "What to wear for an interview tomorrow", "What to wear in this weather"
🎯 What to Do"Don't know what to do this weekend", "What to do after work", "Want to exercise, what should I choose"
🛒 What to Buy"What gift for my friend", "Want to buy headphones, recommend one", "Help me choose something cost-effective"

Quick Commands

CommandFunction
/decide foodQuick food decision
/decide outfitQuick outfit decision
/decide activityQuick activity decision
/decide buy [item]Quick purchase decision
/decide randomCompletely random decision

Core Features

🍽️ Food Decisions

Applicable Scenarios: Breakfast, lunch, dinner, snacks, group dining

Optional Inputs:

  • Budget range (Economy / Standard / Unlimited)
  • Taste preference (Light / Heavy / Spicy / Sweet, etc.)
  • Time constraint (Rushed / Normal / Relaxed)
  • Dietary restrictions (Vegetarian / Halal / Allergies, etc.)

Output:

  • Recommended dish/restaurant
  • Reasoning (taste, nutrition, convenience)
  • Estimated price and time
  • Alternative options

Example:

User: What to eat for lunch? Budget 50, short on time

AI: 🍽️ Lunch Decision: Business Set Meal

📍 Recommendation: [Yoshinoya] Beef Bowl Set
📋 Reasoning: Fast service (3-5 min), within budget (¥45), filling
⏱️ Time: 20-25 minutes

💡 Alternatives: Convenience store bento (faster) / Light salad (healthier)

👔 Outfit Decisions

Applicable Scenarios: Daily wear, occasion dressing, weather adaptation

Optional Inputs:

  • Occasion (Work / Date / Exercise / Formal event)
  • Weather/temperature
  • Style preference (Minimalist / Fashionable / Vintage, etc.)

Output:

  • Complete outfit plan (Top + Bottom + Shoes + Accessories)
  • Styling reasoning (color, occasion fit, comfort)
  • Weather adaptation notes
  • Alternative outfits

Example:

User: First date tomorrow, 18°C, what to wear?

AI: 👔 Date Outfit: Clean Gentleman Style

🧥 Outfit:
• White T-shirt + Light gray casual blazer
• Dark blue slim-fit jeans
• White sneakers + Simple watch

📋 Reasoning: Can remove blazer at 18°C, suitable for restaurant + movie, clean color palette
💡 Alternatives: More formal (shirt + leather shoes) / More casual (remove blazer)

🎯 Activity Decisions

Applicable Scenarios: Weekend activities, after-work plans, exercise choices, work-life balance

Optional Inputs:

  • Available time (duration)
  • Energy level (High / Medium / Low)
  • Budget (Free / Low / Medium / High)
  • Group size (Solo / Friends / Family)

Output:

  • Recommended activity plan
  • Activity reasoning (fit analysis)
  • Required preparation
  • Estimated duration and cost
  • Alternative activities

Example:

User: Weekend alone, don't want to spend much, want to go out

AI: 🎯 Weekend Activity: Park Picnic + Reading

📍 Recommendation: City park picnic
📋 Reasoning: Solo-friendly, low cost (under ¥50), outdoor + digital detox
🎒 Preparation: Picnic mat, sandwich, fruit, book to read
⏱️ Duration: 3-4 hours
💰 Cost: ¥30-50

💡 Upgrade: Bring camera for photography practice / Bird and plant watching

🛒 Purchase Decisions

Applicable Scenarios: Gift shopping, daily necessities, electronics comparison, value shopping

Optional Inputs:

  • Product category
  • Purchase purpose/usage scenario
  • Budget limit
  • Recipient info (age/gender/preferences)
  • Priority (price/quality/brand/features)

Output:

  • Recommended product/brand (1-3 options)
  • Reasoning (fit analysis)
  • Price comparison
  • Purchase channel suggestions

Example:

User: Birthday gift for mom, budget 300-500

AI: 🎁 Gift Decision: Silk Scarf + Hand Cream Set

🎯 Recommendation: Wensli silk scarf + L'Occitane hand cream
📋 Reasoning: Matches 50+ elegant taste, practical, mom values skincare
💰 Budget: ¥300-500
🛒 Channel: Tmall flagship store

💡 Alternatives: Massager (for neck/shoulder pain) / Tea gift set (if she loves tea)

Decision Templates

Standard Output Format

🎯 Decision: [Specific recommendation]

📋 Reasoning:
• [Reason 1 - Preference match]
• [Reason 2 - Scenario fit]
• [Reason 3 - Other value]

💰 Estimated Cost: [Price range]
⏱️ Estimated Time: [Time needed]

💡 Alternatives:
1. [Alternative 1]
2. [Alternative 2]

❓ Satisfaction: [👍 Satisfied / 👎 Not satisfied / 🔄 Try another]

Quick Decision Mode

When user says "hurry", "whatever", "quick", use simplified output:

🚀 Quick Decision: [Result]

Reason: [One sentence]

Say "another one" if not satisfied

Usage Examples

Example 1: Workday Lunch

User: What to eat for lunch? Budget 50, near office, short on time

AI: 🍽️ Lunch Decision: Business Set Meal

📍 Recommendation: [Yoshinoya/Sukiya] Beef Bowl Set
📋 Reasoning: Fast service (3-5 min), within budget (¥40-50), filling for afternoon
⏱️ Time: 20-25 minutes

💡 Alternatives: Convenience store bento (faster) / Light salad (healthier)

Example 2: Date Outfit

User: First date tomorrow night, 18°C, dinner and movie

AI: 👔 Date Outfit: Clean Gentleman Style

🧥 Outfit:
• White T-shirt + Light gray casual blazer
• Dark blue slim-fit jeans
• White sneakers + Simple watch

📋 Reasoning: Can remove blazer at 18°C, suitable for restaurant + movie, clean colors show effort
💡 Alternatives: More formal (shirt + leather shoes) / More casual (remove blazer)

Example 3: Weekend Activity

User: Weekend alone, don't want to spend much, want to go out

AI: 🎯 Weekend Activity: Park Picnic + Reading

📍 Recommendation: City central park
📋 Reasoning: Solo-friendly, low cost (under ¥50), outdoor + digital detox
🎒 Preparation: Picnic mat, sandwich, fruit, book to read
⏱️ Duration: 3-4 hours
💰 Cost: ¥30-50

Example 4: Gift Shopping

User: Birthday gift for mom, budget 300-500

AI: 🎁 Gift Decision: Silk Scarf + Hand Cream Set

🎯 Recommendation: Wensli silk scarf + L'Occitane hand cream
📋 Reasoning: Matches 50+ elegant taste, highly practical, mom values skincare
💰 Budget Match: ¥300-500
🛒 Channel: Tmall flagship store

💡 Alternatives: Massager (for neck/shoulder pain) / Tea gift set (if she loves tea)

Decision Principles

1. Decision First

Give a clear decision result first, then explain the reasoning

2. Transparent Reasoning

Help users understand the logic behind the decision

3. Provide Alternatives

Always provide 1-2 alternative options

4. Feedback Loop

Every output includes a feedback entry point

5. Personalization

Adjust recommendations based on user context and preferences


Safety Escalation (High-Risk Response)

⚠️ Stop the normal decision-support flow immediately if the user expresses any of the following:

  • self-harm or suicide thoughts related to decisions
  • overwhelming despair or inability to make any decisions
  • obvious acute mental-health crisis that makes decision support inappropriate

Use a direct response like:

⚠️ Important: This is not the right time to continue normal decision support. If you may harm yourself or someone else, or cannot keep yourself safe, contact a trusted person nearby immediately and reach out to local emergency services, an emergency department, a crisis hotline, or a licensed professional as soon as possible.

Then keep the tone calm, direct, and serious. Do not continue normal decision guidance until safety is addressed.


Boundaries & Limitations

Supported

✅ Daily trivial matters: food, clothing, activities, shopping

Not Supported

❌ Major life decisions: career choices, investment decisions, medical decisions, legal decisions

Disclaimer

⚠️ Disclaimer: This tool is only for everyday low-stakes decision support and does not constitute medical advice, investment advice, legal advice, or any other professional guidance. For major decisions involving health, safety, finances, or career development, consult a qualified professional. Users remain responsible for their own decisions and outcomes.

This skill can:

  • help with daily trivial decisions (food, clothing, activities, shopping)
  • reduce choice anxiety through quick recommendations
  • provide alternatives for user to choose from

This skill must not:

  • make decisions about mental health treatment
  • claim the user has decision-making disorders
  • present recommendations as medical or psychological advice
  • replace professional counseling, legal advice, or financial planning

Technical Information

AttributeValue
Skill IDdecision-fatigue-reliever
Version1.0.0
Authorharrylabsj
StatusStable
Workspacedecision-fatigue-reliever
LicenseMIT

Last updated: 2026-03-23

相关技能

诊断生产力系统反复失效的根因,给出最小干预——容量测算、瓶颈定位、可靠的本地记录。

作者 Iván855 次安装69 星标

在本地磁盘以分类纯 Markdown 文件保存需要长期留存的事实,与智能体内置记忆并存。

作者 Iván559 次安装18 星标

把自然语言描述转为结构化 JSON,并由 mcp-diagram-generator MCP 服务生成 Draw.io、Mermaid 或 Excalidraw 图表文件。

作者 nssa.io1.0k 次安装47 星标

从 AdMapix API 拉取广告创意、应用、榜单和收入预估等数据,原样返回结构化 JSON。

作者 fly0pants4.3k 次安装296 星标

以 AI 机器人身份加入视频会议,提供语音、虚拟形象与屏幕共享四种模式。

作者 johnpatternai22 次安装8 星标

harrylabsj 的更多技能

浏览全部技能

淘宝与京东、拼多多、唯品会同款比价,输出可执行的购买建议。

作者 harrylabsj52 次安装2 星标

根据美团可见的配送费、包装费、满减门槛、ETA、评价等公开信息,给出一个明确的下单、换店、凑单、再等等或不点的判断。

作者 harrylabsj44 次安装1 星标

在下单前拆解拼多多低价:百亿补贴、拼团门槛、店铺风险、退货条款一次过。

作者 harrylabsj34 次安装1 星标

定位技能瓶颈,把练习日志转化为可执行的下一次训练方案。

作者 harrylabsj35 次安装

用通俗中文解读常规体检和化验报告,标注异常、分清轻重、给出下一步动作。

作者 harrylabsj27 次安装

为盒马鲜生购物规划清单、评估鲜度、推荐配送时段,并整理手动下单步骤。

作者 harrylabsj23 次安装