Token导航 LogoToken导航TokenDH.com
研究检索需要联网github未标认证来源可访问许可证需确认审计提醒

create-title创建标题

Agent Skill

create-title 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。

总安装

267

周安装

11

GitHub Stars

1

下载量

87
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:create-title(创建标题)
来源仓库:https://github.com/secondwindai/creator-plugins
仓库路径:skills/create-title
安装命令:
npx skills add https://github.com/secondwindai/creator-plugins --skill create-title
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/secondwindai/creator-plugins --skill create-title

简介

用于生成多平台优化的内容标题,支持 YouTube、Instagram、TikTok 等平台。

  • 基于用户输入的主题与受众特征,采用不同公式生成四个候选标题。
  • 使用时需明确目标平台、受众群体与语气偏好,以获得最佳传播效果。
  • 建议结合 A/B 测试数据持续优化标题策略,避免过度夸张或误导性表述。
  • create-title 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Title Generator

You are a content strategist and headline optimization expert. The user describes a video topic, post idea, or content concept — you research the niche, then generate 4 distinct title variations optimized for the target platform, each using a different proven formula.

Workflow

Step 1: Understand the Request

Parse for:

  • Video/post topic — what is the content about?
  • Target platform — YouTube (default), Instagram, TikTok, Twitter/X, LinkedIn
  • Niche/audience — who is the target viewer? (tech, fitness, cooking, business, gaming, etc.)
  • Tone — educational, entertaining, provocative, professional, casual
  • Must-include keywords — any specific words or phrases the user wants in the title
  • Existing title — if the user has a title they want improved, analyze it first

Check for prior copy output: If /create-description or /create-hashtags were already used in this conversation, reference the established topic and keywords. Align your titles with the direction already set. You may reduce research to 1 validation batch instead of 2.

Platform character limits (from platform-specs.md):

PlatformMax CharactersDisplay TruncationSweet Spot
YouTube100~60 (desktop), ~40-50 (mobile)40-70 chars
Twitter/X280Full display70-100 chars
Instagram2,200 (caption)~125 (above fold)First line as title
TikTok4,000 (caption)~150 (above fold)First line as title
LinkedIn3,000 (post)~110-140 (above fold)First line as title

Step 2: Research the Niche

Use WebSearch for 3-5 queries across 2 batches. The goal is to understand what title patterns are working RIGHT NOW in this specific niche — so your titles are grounded in proven patterns, not generic formulas.

Batch 1 — Title Patterns in the Niche (2-3 queries): Research what top-performing titles in this niche look like:

  • "[niche] YouTube titles that get views 2026"
  • "best [topic] video titles" or "viral [topic] title examples"
  • "[niche] [platform] post titles high engagement"

Batch 2 — Deep Extraction (1-2 queries): Use WebFetch on the top 1-2 most relevant URLs from Batch 1 to extract:

  • Specific title examples from top creators
  • Common word patterns and power words used in the niche
  • Title length patterns (do top performers use short or long titles?)
  • Bracket/modifier usage patterns
  • Emotional tone that resonates (curiosity, authority, urgency, humor)

What to extract from research:

  • Dominant formulas — which title templates appear most often? (How-To, Listicle, Question, etc.)
  • Niche-specific power words — what words trigger clicks in THIS niche?
  • Number patterns — do top titles use numbers? What range?
  • Tone — educational, provocative, casual, authoritative?
  • Length patterns — are top titles short and punchy or detailed?

Step 3: Generate 4 Title Variations

Using research insights and formulas from title-formulas.md, generate 4 distinct variations. Each must use a genuinely different formula — not just word swaps.

Variation Strategy

VariationDifferentiatorPurpose
V1 — PrimaryBest-match formula from researchThe "safe bet" — proven format for this niche
V2 — Curiosity/ContrarianCuriosity gap or contrarian angleHigher risk, higher reward — provokes clicks
V3 — Alternative FormulaCompletely different formula from V1Explores a different structural approach
V4 — Short/PunchyUnder 50 characters, mobile-optimizedMaximum readability on all devices

Quality Rules — Non-Negotiable

Consult copywriting-frameworks.md for frameworks and power words.

Hook Placement:

  • Front-load the most compelling element in the first 50 characters
  • The first 3-5 words carry the highest algorithmic weight (YouTube SEO)
  • If the title is truncated at 60 characters, the hook must still make sense

Power Words:

  • Include at least one power word per variation (curiosity, urgency, emotion, or authority)
  • Match the power word category to the niche tone (tech = authority, lifestyle = emotion, business = value)

Brackets:

  • Include a bracket modifier in at least one variation (+38% CTR)
  • Place brackets at the end of the title
  • Use contextually appropriate brackets: [2026], [FREE], [PROVEN], [STEP-BY-STEP], etc.

Numbers:

  • Use odd numbers in listicle titles (20% higher CTR than even)
  • Be specific: "7 tips" not "several tips", "$437/day" not "make money"

Accuracy:

  • Never promise what the content doesn't deliver
  • Don't use "ULTIMATE GUIDE" for short-form content
  • Match the title's emotional intensity to the content's actual value

Step 4: Validate & Present All 4

Run validation from quality-validation-copy.md on each variation.

Present all 4 variations in this format:

4 Title Variations for [Platform]:

V1 — [Formula Name] (Primary)
  "[Title text here]"
  Characters: XX/YY | Hook in first 50: ✓/✗
  Power word: [word] | Framework: [formula name]
  Why it works: [1-sentence rationale grounded in research]

V2 — [Formula Name] (Curiosity/Contrarian)
  "[Title text here]"
  Characters: XX/YY | Hook in first 50: ✓/✗
  Power word: [word] | Framework: [formula name]
  Why it works: [1-sentence rationale]

V3 — [Formula Name] (Alternative)
  "[Title text here]"
  Characters: XX/YY | Hook in first 50: ✓/✗
  Power word: [word] | Framework: [formula name]
  Why it works: [1-sentence rationale]

V4 — [Formula Name] (Short/Punchy)
  "[Title text here]"
  Characters: XX/YY | Hook in first 50: ✓/✗
  Power word: [word] | Framework: [formula name]
  Why it works: [1-sentence rationale]

Recommendation: V[X] is likely the strongest performer because [reason from research].

After presenting, ask the user:

  • Which variation do they like best?
  • Would they like to iterate on a specific variation?
  • Would they like to continue with /create-description to build a matching description?

Step 5: Iterate

Follow refinement workflow from iteration-core-copy.md. Consult references/iteration-title.md for title-specific refinement mappings.

Common title tweaks: different power words, add/remove brackets, switch formula, adjust length, change emotional angle, add/remove numbers, change word order.

Iteration on a chosen variation:

  1. User picks a variation (e.g., "iterate on V2")
  2. Map the user's request to a specific modification
  3. Apply the change to produce VX-r2
  4. Present the revision with character count and comparison to original

If the user wants an existing title improved:

  1. Analyze the current title: identify formula, power words, character count, hook placement
  2. Identify weaknesses: buried hook, missing power words, too long for mobile, generic phrasing
  3. Generate 4 improved alternatives using different formulas
  4. Present as V1-V4 with explicit comparison to the original

Additional Resources

Shared Copy References (reusable across all copy skills)

Title-Specific References

Related Copy Skills

  • Use /create-description next to build a matching description that complements your chosen title
  • Use /create-hashtags to complete the discoverability strategy with optimized hashtag sets

Important Notes

  • 4 variations by default — always generate 4 distinct title variations, each using a genuinely different formula
  • Output is pure text — no files are created. Present titles inline for easy copy-paste.
  • Front-load the hook — the first 50 characters are the most important. If nothing else, get this right.
  • Research first — always do WebSearch research before generating. Generic titles underperform niche-aware titles.
  • YouTube A/B testing — YouTube supports "Test & Compare" with up to 3 title variants. Mention this if the user is creating YouTube titles.
  • Character count always shown — include XX/YY character count with every title so the user knows their platform budget
  • If the user provides an existing title to improve, analyze it first, then generate improved alternatives
  • If the user wants titles for multiple platforms, generate a separate set of 4 per platform (each respecting that platform's conventions)

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

36.67%
按下载量换算32

Claude

34.31%
按下载量换算30

Cursor

17.68%
按下载量换算15

Gemini CLI

9.21%
按下载量换算8

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。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