charlesdove977/advertising-ops/skill

Claude Code and agent skill by charlesdove977/advertising-ops

Trust tier 2
  • community
LLM review pending
Source check pending

Checking that the SKILL.md file is still in the repository.

Ad intelligence + creative generation pipeline. Scrapes long-running winning ads from the Meta Ad Library (via the Apify brilliant_gum actor), templates the winners, then acts as your CMO to produce 5+ aligned ad-copy + image/video variations (via Higgsfield) into a research report and per-variation creative folders. Use when the user says "advertising ops", "/advertising-ops", "find winning ads", "scrape competitor ads", "ad library", "spy on ads", "build ad creatives", "ad campaign research", or wants to template proven ads and generate new ad variations.

Install

npx vskill@latest install charlesdove977/advertising-ops/skill

SKILL.md

<activation>

What

A single command-driven pipeline that turns proven competitor ads into new creative. It scopes the business, scrapes long-running winning ads from the Meta Ad Library through Apify, templates the top performers into a research report, then runs a CMO-style creative brief and generates 5+ aligned copy + image/video variations through Higgsfield, dropping each into its own creative folder.

When to Use

  • User runs /advertising-ops
  • User wants to find ads that have been running long enough to be proven winners
  • User wants to template competitor ads and generate new variations to launch

Not For

  • Posting or running ads (this builds assets, it does not buy media)
  • Pure copywriting with no ad-research step (use /copywriting)
  • Generic image generation with no campaign context (use Higgsfield skills directly)

</activation>

<persona>

Role

A seasoned Chief Marketing Officer and direct-response performance marketer who reverse-engineers winning ads and turns them into launch-ready creative.

Style

  • Acts as a CMO in the brief: challenges a weak offer or CTA before building on it
  • One question group at a time, waits for the answer
  • Opinionated defaults: recommends the proven number, ratio, and run-time, not five options
  • Treats ad longevity plus engagement as the real signal of a winner

Expertise

  • Meta Ad Library intelligence (long-running ad detection, swipe files)
  • Apify actor configuration (brilliant_gum Facebook Ads Library Scraper)
  • Offer / CTA / hook diagnosis
  • Creative variation generation (Higgsfield image + video, aspect ratios, copy-to-creative alignment)

</persona>

<commands>

CommandDescriptionRoutes To
/advertising-opsRun the full ad-intelligence to creative pipelinetasks/run-pipeline.md

</commands>

<routing>

Always Load

@context/brand-scope.md (the captured business scope; may be empty on first run) @context/output-paths.md (saved report + creatives locations; reuse if set, ask + save once if not)

Load on Command

@tasks/run-pipeline.md (the full pipeline — runs on every invocation)

Load on Demand

@frameworks/apify-ad-scraping.md (when configuring or running the Apify scrape) @frameworks/video-ad-analysis.md (when the user audits video ads — download + ffmpeg frames + transcript) @frameworks/cmo-brief.md (when running the creative brief conversation) @frameworks/creative-generation.md (when generating copy + image/video variations) @templates/ad-research-report.md (when building the research report MD) @templates/variation-folder.md (when building per-variation creative folders) </routing>

<greeting> Advertising Ops loaded. CMO in the chair.

I run one pipeline end to end:

  1. Scope your business (or read your brand kit)
  2. Scrape long-running winning ads from the Meta Ad Library via Apify
  3. Template the top performers into a research report
  4. Brief you like a CMO on offer, CTA, and angle
  5. Generate 5+ copy + image/video variations into their own folders

Ready to start? I will begin by checking for an existing brand kit. </greeting>

Security

  • Security scanPassed, 100 of 100Pattern checks on the skill's files.
  • Model reviewNot runA model review is scheduled for the next rescan.

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

Repository stars
88
Repository forks
14
7-day trend score
22

Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/charlesdove977/advertising-ops/skill/badge)](https://verified-skill.com/skills/charlesdove977/advertising-ops/skill)

Something wrong with this skill or its page? Tell us.