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

meta-ads-creative元广告创意

Agent Skill

meta-ads-creative 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

4,375

周安装

177

GitHub Stars

4

下载量

1,374
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:meta-ads-creative(元广告创意)
来源仓库:https://github.com/cdeistopened/opened-vault
仓库路径:skills/meta-ads-creative
安装命令:
npx skills add https://github.com/cdeistopened/opened-vault --skill meta-ads-creative
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/cdeistopened/opened-vault --skill meta-ads-creative

简介

meta-ads-creative 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中快速定位候选结果。

  • 适用于关键词搜索、任务场景匹配或来源线索梳理等研究检索类工作。
  • 通过 npx skills add 命令从指定 GitHub 仓库安装并使用该技能。
  • 安装前需确认权限范围、维护状态,注意是否涉及联网、命令执行或文件读写操作。
  • 适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

SKILL.md

Meta Ads Creative Skill

Purpose

Create Meta ad creative that converts using research-driven development, the 6 Elements framework, and format fitting - matching your message to proven ad formats.

Core Philosophy: The best Meta ads don't look like ads. Lo-fi > polished. Authentic > produced. Native > interruptive.

When to Use

  • Create ad creative for Meta campaigns
  • Develop concepts for retargeting or cold audiences
  • Generate multiple variations for A/B testing
  • Transform testimonials into ad creative

Related Skills:

  • hook-and-headline-writing - Optimizing Primary Text hooks
  • opened-identity - Brand voice alignment
  • social-content-creation - Organic content that can be boosted

The Value Formula

Value = Dream Outcome x Perceived Likelihood / Time Delay x Effort

Every ad must communicate: (1) What transformation they want, (2) Will it work for them, (3) How long until results, (4) How hard is it.


The 6 Elements of Ad Creative

Every Meta ad has 6 elements:

ElementPurposeKey Guidance
MediaStop the scroll (80% of performance)iPhone footage > professional. Faces > stock.
Primary TextHook + body copy above mediaFirst 125 chars show before "See More"
HeadlineShort promise below media5-8 words max
DescriptionSecondary text below headlineUrgency, proof, or clarification
CTA ButtonAction button"Learn More" (cold) / "Apply Now" (warm)
PageLanding destinationMessage must match ad
Load on-demand: references/6-elements-framework.md for detailed guidance on each element.

The 4-Phase Workflow

Phase 1: Research

Goal: Understand audience language before writing.

Sources to mine:

  • Testimonials and reviews (exact language)
  • Support tickets and FAQs (objections)
  • Reddit/Facebook groups (common questions)
  • Competitor ads running 30+ days (proven angles)
  • Top organic posts (what resonates)

Output: [Campaign]_Research.md with 10-20 insights and exact quotes.

Load on-demand: references/creative-research-methods.md for detailed research process. Load on-demand: references/audience-segments.md for OpenEd segment mapping.

Phase 2: Copywriting

Goal: Transform research into compelling copy using proven formulas.

Match Awareness Level

LevelAudience StateCopy Approach
1 - UnawareDon't know problem existsLead with problem agitation
2 - Problem AwareKnow problem, not solutionsValidate + introduce solution
3 - Solution AwareKnow solutions, not youDifferentiate your approach
4 - Product AwareKnow you, haven't actedAddress objections, provide proof
5 - Fully AwareReady to actMake offer irresistible

OpenEd typical: Cold traffic = Level 1-2. Retargeted = Level 3-4. Email list = Level 4-5.

Write Hooks First

The hook is 80% of performance. Types: Stated (what you SAY), Visual (what they SEE), Audio (what they HEAR).

Load on-demand: hook-and-headline-writing skill has 15+ proven hook formulas with examples.

Body Copy Formulas

  • PAS - Problem, Agitate, Solution
  • AIDA - Attention, Interest, Desire, Action
  • Before-After-Bridge - Current state, desired state, how to get there
  • Testimonial - Quote, context, outcome, invitation

Output: [Campaign]_Copy_Variations.md with 5-10 variations per concept.

Load on-demand: references/copywriting-formulas.md for detailed formulas with examples.

Phase 3: Format Selection

Goal: Match copy to the format that amplifies it best.

Lo-Fi Native Formats (High Performance)

FormatWhy It WorksBest For
Notes AppLooks like personal contentProblem-solution, starter packs
Text-Over-VideoStory through text sequenceTransformations, permission
Reddit/Tweet ScreenshotDiscovery energyHot takes, personal stories
Instagram CommentDialogue/Q&A feelAddressing objections
Meme FormatsCulturally nativeContrasts, humor
Testimonial CardDirect social proofWarm audiences

Format-Audience Matrix

Audience TempBest Formats
ColdNotes App, Meme, Text-Over-Video, Reddit/Tweet
WarmUGC, Testimonial, Carousel, Before/After
HotTalking Head, Demo, Direct Offer

Output: Select 3-5 formats per campaign.

Load on-demand: references/ad-formats-library.md for all formats with templates and examples.

Phase 4: Assembly & Testing

Goal: Combine copy + format into complete concepts, then generate variations.

Document Each Concept

## Ad Concept: [Name]

**Format:** [Selected format]
**Audience:** [Segment] | **Awareness:** [Level 1-5]

**Media:** [Description]
**Primary Text:** [Full copy]
**Headline:** [5-8 words]
**Description:** [Secondary line]
**CTA:** [Button choice]
**Page:** [Landing destination]

Testing Variations

  • Hook Testing: Same format, different opening lines (3-5 versions)
  • Format Testing: Same message, different formats (3 versions)
  • Angle Testing: Same format, different emotional angles (fear vs hope)

Naming Convention

[Campaign]_[Format]_[Hook Type]_[Variation] Example: OpenEd_NotesApp_StarterPack_V1

Output: [Campaign]_Ad_Concepts.md with complete concepts.


Quality Checklist

Before finalizing any concept:

Copy:

  • Hook stops scroll in first 3 seconds
  • Value formula addressed
  • Awareness level matched
  • Sounds like real person, not corporate

Visual:

  • Lo-fi/native feel
  • Text readable on mobile
  • Authentic, not stock

Strategic:

  • Landing page message matches
  • Clear next step for viewer

OpenEd-Specific Notes

Brand Voice: Warm, empowering, not salesy. "We're here to help" not "Buy now."

Key Messages:

  • Tuition-free access to resources
  • Maintain educational freedom
  • Support without judgment

Testimonials: Use real quotes verbatim. Match to audience segment. Get permission.

Compliance: No outcome guarantees. Follow Meta ad policies.


References (Load On-Demand)

ReferenceContents
references/6-elements-framework.mdDetailed element guidance
references/ad-formats-library.mdAll formats with templates
references/copywriting-formulas.mdPAS, AIDA, hooks, headlines
references/creative-research-methods.mdResearch process
references/audience-segments.mdOpenEd segment mapping
examples/Existing OpenEd concepts

Related Skills

  • dude-with-sign-writer - One-liners for text-only ads (12 punchy patterns)
  • hook-and-headline-writing - Hook formulas (extensive overlap, use for hook optimization)
  • social-content-creation - Organic content that can become ads
  • opened-identity - Brand voice framework
  • ghostwriter - Authentic voice guidelines

Vault references: References/testimonials.md (parent quotes) and References/success-stories.md (25+ student stories) for social proof in ad creative.

*v1.1 (2026-01-23): Refactored for progressive disclosure. Detailed content moved to references.*

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

34.23%
按下载量换算470

Claude

32.73%
按下载量换算450

Cursor

17.64%
按下载量换算242

Gemini CLI

10%
按下载量换算137

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

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