Back to Blog
March 7, 2026
Open SourceSkillsAmp

The Most Comprehensive Open-Source AI Skills for Advertising

36 professional-grade skills for every ad platform, free and MIT-licensed.

Why We Built This

We just open-sourced the largest collection of AI agent skills for advertising. 36 professional-grade skills covering attribution modeling, PMax optimization, creative testing, budget planning, and every major ad platform.

AI coding agents like Amp are transforming how developers work. But when it comes to advertising, there is a massive knowledge gap. Generic LLMs know marketing theory. They do not know that LinkedIn's API v202510 requires politicalIntent as a string enum, or that PMax needs a bulk mutate request with all assets linked atomically, or that Reddit's v3 API uses flat string arrays for geolocations.

Skills bridge that gap. Each one is a structured markdown file that injects deep, platform-specific knowledge directly into the agent's context. Real GAQL queries, API payloads, platform mechanics, and battle-tested workflows.

What Makes This Collection Different

There are other skill repositories out there. Most are generic productivity tools or simple prompt templates. Here is why this collection stands apart:

Platform-Specific Depth

Not generic “marketing tips.” Real GAQL queries, Rest.li 2.0 payloads, Reddit v3 API quirks, and Meta Graph API patterns. The kind of knowledge that prevents 3 AM debugging sessions.

Built by Practitioners

Every skill comes from managing real ad spend across Google, Meta, LinkedIn, Amazon, TikTok, Reddit, and X. These are not theoretical. They are extracted from production campaigns.

Interconnected

Skills reference each other. The budget optimizer feeds into the ROAS calculator, which connects to the dayparting scheduler. Run them together for compound insight.

Actionable Formulas

Python code, SQL queries, statistical tests, and calculation templates you can use immediately. Not “consider optimizing” but actual formulas with real numbers.

How Skills Work

Skills are instruction files (SKILL.md) that teach AI coding agents how to perform specific tasks. Each skill includes:

  • Frontmatter with name and description for auto-matching
  • Step-by-step workflows the agent follows
  • Reference data including formulas, benchmarks, and industry data
  • Examples with sample commands, scripts, and expected outputs

Drop a skill into your .agents/skills/<name>/SKILL.md folder and Amp automatically loads it when your request matches the skill's description.

your-project/
  .agents/
    skills/
      attribution-modeling/
        SKILL.md
      meta-ads-diagnostics/
        SKILL.md
      performance-max-optimizer/
        SKILL.md
      ... (36 skills)

The Full Catalog

36 skills across 10 categories. Every one of them free, MIT-licensed, and ready to use.

Measurement and Attribution

attribution-modeling Multi-touch models: first-touch, last-touch, linear, time-decay, position-based, data-driven, Markov chain

incrementality-testing Geo holdout tests, conversion lift studies, ghost ads, statistical significance

roas-calculator ROAS, CPA, and campaign profitability metrics

anomaly-detector Z-score and IQR detection for spend spikes and CTR drops

Platform-Specific Optimization

meta-ads-diagnostics Learning Phase, Breakdown Effect, creative fatigue, Advantage+

performance-max-optimizer Asset groups, search term insights, audience signals, channel allocation

google-ads-quality-score Expected CTR, Ad Relevance, Landing Page Experience with GAQL

google-shopping-optimizer Feed optimization, Merchant Center diagnostics, custom labels

linkedin-ads-targeting B2B strategy, ABM setup, Lead Gen Forms, cost benchmarks

amazon-ads-optimizer Sponsored Products/Brands/Display, ACOS/TACOS analysis

Creative and Content

ad-copy-generation RSA headlines, descriptions, and ad variations for Google, Meta, LinkedIn

ugc-creative-brief Hook formulas, shot lists, platform specs, AI tool comparison

video-ad-scriptwriter 6s bumpers, 15s pre-roll, 30s narrative, TikTok native hooks

ad-creative-fatigue-detector Frequency vs CTR decay analysis, creative lifespan prediction

creative-testing Cross-platform A/B testing with statistical significance

recording-demo-videos Playwright-based recording with zoom effects, text overlays, smooth scrolling

Budget and Planning

budget-optimizer Cross-platform budget reallocation based on ROAS across 9+ platforms

bid-optimization CPC adjustments, day-parting modifiers, audience bid strategies

media-plan-builder Channel mix, CPM/CPC projections, reach estimates

dayparting-scheduler Hour-of-day heat maps and bid modifier calculations

seasonal-budget-planner BFCM scaling, CPM inflation forecasting

Account Structure and Hygiene

campaign-structure-auditor SKAG vs STAG audit, duplicate detection, naming conventions

negative-keyword-miner Search term analysis with AI categorization and industry templates

campaign-preflight Pre-launch checks for geo targeting, exclusion lists, tracking setup

Research and Analysis

keyword-research High-intent keyword discovery via Google Keyword Planner with competition analysis

competitor-analysis Competitor ad analysis using Facebook Ads Library and auction insights

audience-expansion-strategy Lookalike audiences, seed optimization, audience layering

Tracking and Data

pixel-capi-auditor Meta Pixel + CAPI, Google Enhanced Conversions, TikTok/LinkedIn/Reddit server-side tracking

utm-builder Bulk UTM generation with naming conventions

first-party-data-strategy Enhanced Conversions, server-side tracking, Consent Mode v2, CDPs

landing-page-optimizer Above-the-fold optimization, message match scoring, form design

Reporting and Compliance

multi-channel-reporting Cross-channel performance reports and executive summaries

executive-reporting C-suite templates with narrative insights, YoY/MoM comparisons

ad-policy-compliance Platform policy checklists for Google, Meta, LinkedIn, TikTok, Reddit

Campaign Execution

cross-platform-launcher Deploy campaigns to multiple ad platforms simultaneously

retargeting-sequence-designer Multi-stage retargeting funnels with sequential messaging and frequency caps

Installation

Three ways to start using these skills today:

Option 1: Copy individual skills

cp -r skills/attribution-modeling \
  your-project/.agents/skills/attribution-modeling

Option 2: Copy all 36 skills

cp -r skills/* your-project/.agents/skills/

Option 3: Use in Synter (no setup needed)

All skills are built into Synter. Sign up and start asking the agent to audit your account structure, build a media plan, or calculate your ROAS.

Project-Based Skill Scoping

Inside Synter, skills go further. You can assign specific skills to specific projects so the agent only loads what is relevant.

Running a Google Ads campaign for an e-commerce client? Assign Google Ads, Google Shopping Optimizer, and Performance Max skills to that project. Running LinkedIn lead gen for a B2B SaaS? Assign LinkedIn Ads Targeting and Lead Gen Forms.

The agent loads only the skills scoped to your current project. This keeps context lean, responses faster, and token costs lower. You can also upload custom skills with your team's naming conventions, approval workflows, and client-specific rules.

Contribute

Have a useful advertising or marketing skill? Open a PR. Each skill should include a SKILL.md with frontmatter, clear workflows, examples, and no hardcoded credentials.

github.com/Synter-Media-AI/free-skills

36 skills · MIT License · PRs welcome

Not using Synter yet? Sign up for free and get 200 credits to try it out.

Ready to let AI agents run your campaigns?

Start for free with 200 credits and launch campaigns across Google, Meta, LinkedIn, Reddit, and more.

The Most Comprehensive Open-Source AI Skills for Advertising | Synter