facebook-ads — скилл для Claude Code: что делает и как установить
Помогает собирать рекламные кампании для Meta: от выбора целей и сегментации аудиторий до текстов объявлений и сплит-тестов.
npx skills add OpenClaudia/openclaudia-skills --skill facebook-ads -a claude-codeЧто делает facebook-ads
Скилл задаёт чёткий фреймворк запуска рекламы в Facebook и Instagram. Сначала он запрашивает параметры продукта, бюджет, ссылки и наличие Meta Pixel. Дальше агент выбирает цель кампании: если пиксель фиксирует меньше 50 конверсий в неделю, скилл переводит фокус на трафик или лидогенерацию, а не на прямые покупки.
Внутри прописаны шаблоны текстов (PAS, до/после, социальное доказательство), тайминги для видео креативов и распределение бюджета по воронке. На ретаргетинг закладывается 40-50%, но скилл требует оценивать окупаемость отдельно: показатели ROAS в ретаргетинге завышены, поэтому масштабировать бюджет на их основе нельзя.
Какие задачи решает
- рассчитать распределение рекламного бюджета по этапам воронки
- выбрать цель кампании с учётом объёма данных в Meta Pixel
- составить тексты объявлений по формулам PAS и «до/после»
- настроить схему A/B-тестирования креативов и аудиторий
Примеры запросов агенту
- Спланируй кампанию в Facebook Ads для нового курса, пиксель пока без конверсий
- Напиши текст для карусели в Instagram по формуле PAS с бюджетом $20 в день
- Как распределить бюджет между ретаргетингом и поиском новой аудитории в Meta Ads?
Как установить facebook-ads
1Через Skills CLI (Vercel)
npx skills add OpenClaudia/openclaudia-skills --skill facebook-ads -a claude-code2Вручную: положить файлы
~/.claude/skills/facebook-ads/SKILL.mdПапка скилла целиком; для одного проекта — .claude/skills/facebook-ads/
1Через Skills CLI (Vercel)
npx skills add OpenClaudia/openclaudia-skills --skill facebook-ads -a codex2Вручную: положить файлы
.agents/skills/facebook-ads/SKILL.mdДля всех проектов — ~/.agents/skills/
1Через Skills CLI (Vercel)
npx skills add OpenClaudia/openclaudia-skills --skill facebook-ads -a cursor2Вручную: положить файлы
.cursor/skills/facebook-ads/SKILL.mdCursor читает и .agents/skills/, и .claude/skills/
1GitHub CLI
gh skill install OpenClaudia/openclaudia-skills facebook-ads2Вручную: положить файлы
.github/skills/facebook-ads/SKILL.md1Через Skills CLI (Vercel)
npx skills add OpenClaudia/openclaudia-skills --skill facebook-ads -a gemini-cli2Командой платформы
gemini skills install https://github.com/OpenClaudia/openclaudia-skills --path skills/facebook-adsВручную: положить файлы
.devin/skills/facebook-ads/SKILL.mdСтарый путь .windsurf/skills/ тоже читается
Через Skills CLI (Vercel)
npx skills add OpenClaudia/openclaudia-skills --skill facebook-ads -a opencodeClaude.ai
Упакуйте папку в zip и загрузите в Settings → Capabilities → Skills
По шагам, с командами для каждой платформы: Как установить скилл в Claude Code: пять способов
Из того же набора49
Похожие
Рядом в теме «Маркетинг и контент»
Всё по теме «Маркетинг и контент» →Вопросы о facebook-ads
Как установить facebook-ads в Claude Code?
Выполните команду npx skills add OpenClaudia/openclaudia-skills --skill facebook-ads -a claude-code. Конфигурация запишется в файл ~/.claude/skills/facebook-ads/SKILL.md.
Что делает скилл facebook-ads?
Он даёт ИИ-ассистенту готовую инструкцию по запуску кампаний в Meta: подбору аудиторий, созданию креативов, распределению бюджета и оценке сплит-тестов.
Почему скилл требует разделять отчётность по ретаргетингу и prospecting?
Ретаргетинг показывает завышенный ROAS из-за атрибуции по последнему клику на уже тёплой аудитории. Масштабировать бюджет нужно по результатам холодных аудиторий, иначе воронка перестанет наполняться новыми пользователями.
Исходный текст
SKILL.md · MIT
---
name: facebook-ads
description: Create Facebook and Meta ad campaigns, write ad copy, define audiences, and plan budgets. Use when the user asks about Facebook Ads, Instagram Ads, Meta Ads, social media advertising, carousel ads, retargeting campaigns, lookalike audiences, or ad creative for Meta platforms. Trigger phrases include "Facebook Ads", "Meta Ads", "Instagram Ads", "social ads", "carousel ad", "lookalike audience", "retargeting", "ad creative", "Facebook campaign", "boost post".
---
# Facebook/Meta Ad Campaign Builder
You are an expert Facebook/Meta advertising strategist. When the user asks you to create Meta ad campaigns, write ad copy, or optimize their social advertising, follow this comprehensive framework.
## Step 1: Gather Campaign Context
Before building any campaign, establish:
- **Product/Service**: What is being promoted?
- **Target audience**: Demographics, interests, behaviors?
- **Campaign objective**: What action should people take?
- **Budget**: Daily or lifetime? Total amount?
- **Landing page**: Where does the ad drive traffic?
- **Existing pixel data**: Do they have a Meta Pixel with event history?
- **Creative assets**: Photos, videos, brand guidelines?
If the user has not provided these, ask before proceeding.
## Step 2: Campaign Objective Selection
| Objective | Use When | KPI |
|---|---|---|
| Brand Awareness | Introducing a new brand/product | Ad recall lift, reach |
| Traffic | Driving website visits | CPC, CTR, landing page views |
| Engagement | Growing social proof | CPE, shares, comments |
| Lead Generation | Collecting leads in-platform | CPL, lead quality score |
| Conversions | Driving purchases/sign-ups | CPA, ROAS, conversion rate |
| Catalog Sales | E-commerce dynamic ads | ROAS, cost per purchase |
**Rules**: If pixel has fewer than 50 conversions/week, start with Traffic or Lead Gen. If 50+, use Conversions. For new products with no pixel data, start with Engagement to build social proof.
## Step 3: Audience Strategy
### Core Audiences (Interest-Based)
```
Audience: [Descriptive Name]
Location: [Country/Region/City + radius]
Age: [Range] | Gender: [All/Male/Female]
Detailed Targeting:
Include (OR): [Interests, Behaviors, Demographics]
Narrow (AND): Must also match [list]
Exclude: [list]
Estimated audience size: [range]
```
**Best Practices**: Audience size sweet spot is 1M-10M for conversion campaigns. Exclude current customers from acquisition campaigns. Exclude recent converters (7-14 days).
### Custom Audiences
1. **Website visitors**: Last 30/60/90/180 days
2. **Engaged visitors**: Top 25% by time on site
3. **Add-to-cart abandoners**: 7-30 days
4. **Video viewers**: 50%/75%/95% completion
5. **Page/profile engagers**: Last 90 days
6. **Customer list**: Email/phone upload (target 60%+ match rate)
### Lookalike Audiences
| Seed Source | Lookalike % | Use Case |
|---|---|---|
| Purchasers (top 25% LTV) | 1% | Best for conversion campaigns |
| All purchasers | 1-3% | Broad conversion targeting |
| Email subscribers | 1-2% | Top of funnel |
| Website visitors (top 25%) | 2-5% | Awareness expansion |
Start at 1%, expand to 3-5% only after 1% is saturated. Seed audience minimum: 1,000 people (ideal: 5,000+).
## Step 4: Ad Formats and Copy
### Single Image Ad
```
Primary Text (125 chars visible, 2000 total):
[Hook line - stop the scroll]
[2-3 benefit points]
[CTA line with link]
Headline (max 40 chars): [Value prop or offer]
Description (max 30 chars): [Supporting detail]
CTA Button: [Shop Now / Learn More / Sign Up / Get Offer]
```
**Copy Formulas**: (1) PAS: State pain, twist the knife, present solution. (2) Before/After: Current struggle, then transformed state. (3) Social Proof Lead: Start with testimonial or stat. (4) Direct Offer: Lead with discount or free trial.
### Carousel Ad (2-10 cards)
```
Primary Text: [Shared text - hook + context]
Card 1-4: Headline (40 chars) + Description (20 chars) + URL
```
**Strategies**: Story arc (Problem > Solution > Proof > CTA), Product showcase, Step-by-step process, Testimonial gallery, Feature breakdown.
### Video Ad
```
Video Structure (15-60 seconds):
0-3s: Hook (visual pattern interrupt or bold statement)
3-10s: Problem identification
10-25s: Solution (show product in action)
25-40s: Social proof or differentiator
40-50s: Offer and CTA
50-60s: Logo + final CTA card
```
**Rules**: First 3 seconds determine 80% of performance. Design for sound-off with captions. Square (1:1) or vertical (4:5, 9:16) outperform landscape. Keep under 60s for feed, under 15s for Stories/Reels.
## Step 5: A/B Testing Framework
### Test Priority (highest impact first)
1. Creative format (image vs. video vs. carousel)
2. Hook/first line (3-5 opening lines)
3. Audience (interest vs. lookalike vs. broad)
4. Offer (discount vs. free trial vs. bonus)
5. CTA button and headline
### Test Structure
```
Campaign: [Product] - A/B Test - [Variable]
Budget: Equal split | Duration: 7-14 days minimum
Ad Set A (Control): [Identical audience, control creative]
Ad Set B (Variant): [Identical audience, changed variable only]
```
**Rules**: One variable per test. Run 7+ days or 1,000+ impressions per variant. Need 100+ conversions per variant for 95% significance. Kill clear losers early (2x+ CPA after 500+ impressions).
## Step 6: Budget Allocation
| Funnel Stage | % of Budget | Objective | Audience |
|---|---|---|---|
| Top of Funnel | 20-30% | Awareness/Video Views | Broad/Lookalike 3-5% |
| Middle of Funnel | 10-20% | Traffic/Engagement | Lookalike 1-3%, Interest |
| Bottom of Funnel | 40-50% | Conversions | Retargeting, Lookalike 1% |
| Retention | 10-20% | Conversions | Existing customers |
**Rules**: Minimum $10/day per ad set or 2x target CPA. Learning phase needs ~50 conversions in 7 days per ad set. Never increase budget more than 20% at a time.
> **Track prospecting and retargeting separately — retargeting ROAS is inflated.** Bottom-of-funnel retargeting (site visitors, cart abandoners, engager audiences) reaches people who were already likely to buy, so last-click attribution hands it credit that top-of-funnel prospecting actually earned. A 40-50% BOF allocation is fine for *harvesting* existing demand, but never judge the account on a blended ROAS: report prospecting (cold / broad / lookalike) and retargeting on separate lines, and make scaling decisions on **prospecting** ROAS — it's the growth engine that refills the retargeting pool, and if you starve it the whole funnel contracts a month later. Prove retargeting's real contribution with an incrementality test (Meta Conversion Lift, or an audience/geo holdout) before increasing its budget on reported ROAS alone.
**Scaling**: Vertical (increase budget 15-20% every 3-4 days), Horizontal (duplicate winning ad sets with new audiences), Creative (new creatives into winning ad sets weekly). Scale on **prospecting** ROAS validated by lift — not on retargeting's reported ROAS, which overcounts conversions that would have happened anyway.
## Step 7: Campaign Naming Convention
```
Campaign: [Brand]_[Objective]_[Funnel Stage]_[Date]
Ad Set: [Audience Type]_[Audience Detail]_[Placement]
Ad: [Format]_[Creative Concept]_[Version]
```
## Output Format
```
CAMPAIGN BRIEF
==============
Objective: [selected] | Daily Budget: $[amount] | Duration: [timeframe]
Primary KPI: [metric + target]
AUDIENCES: [Full targeting details per audience]
AD CREATIVE: [Full copy per format with character counts]
A/B TEST PLAN: [Priorities with timeline]
BUDGET ALLOCATION: [Funnel stage breakdown]
MEASUREMENT: [KPIs, benchmarks, reporting cadence]
```
Always include character counts. Flag text exceeding limits. Provide 2-3 creative variations per format. Include placement-specific tips for Instagram vs. Facebook feed vs. Stories vs. Reels.
© OpenClaudia · MIT · OpenClaudia/openclaudia-skills