Claude Ads
Claude Code skill that audits paid ad accounts across 12 platforms and reports what to fix first.
- Stars
- 10.7k
- Forks
- 1.6k
- Licence
- MIT
- Language
- Python
- Last update
- 4 days ago
- Type
- AI model / agent
What it is
Claude Ads is a free, MIT-licensed skill for Claude Code. You give it ad account data such as exports, screenshots, or summaries, and it reports what is working, what is broken, and what to fix first across twelve ad platforms. The README says it reads data and reports back, and does not change accounts on its own. Every check ends as pass, fail, unknown, or not applicable, so missing data is shown rather than guessed.
- Run /ads audit google or /ads audit all on exported account data to list passed, failed, and missing checks
- Use /ads plan to build channel, budget, measurement, and competitor plans
- Use /ads create to draft copy, image, video, or product-photo briefs
- Use /ads monitor to review pacing, tracking, ad fatigue, and policy issues
- Use /ads report to render a saved run again as Markdown, HTML, or PDF
Great for
- Agencies and consultants reviewing several client ad accounts
- In-house marketing teams that want a structured audit of their paid campaigns
- People already using Claude Code who want ad-account reviews inside it
- Teams that want reports saved as versioned JSON, Markdown, HTML, or PDF
Not ideal for
- Anyone wanting it to make live changes to ad accounts automatically, since the README says it does not change accounts on its own
- Users without Claude Code or a compatible agent host, since it is built for that runtime
- Teams needing verified support beyond Claude Code, since the README marks other hosts as experimental
How to get it
- 1
Install with a package manager
You'll need: Claude Code · CPython 3.11 or 3.12
Add the marketplace in Claude Code
/plugin marketplace add agricidaniel/claude-adsInstall the plugin in Claude Code
/plugin install claude-ads@ai-marketing-hub-claude-ads - 2
Run it
- Run /ads setup and answer the questions about your business, goal, and approvers; never paste API keys or tokens
- Provide evidence such as a CSV or JSON account export, screenshots, or a summary with date range, currency, and timezone
- Run /ads audit google or /ads audit all
- Run /ads next to see the most important blocker
- 3
Get the most out of it
- Use the --draft flag on /ads optimize and /ads launch to review changes before anything is applied
- Use /ads validate to check contracts, runs, capabilities, or release readiness
- Use /ads research refresh to update platform, policy, and API evidence
- Try the optional Cockpit workbench with its built-in synthetic demo to review finished audits
Downloads
Strengths
- Covers twelve ad platforms with a per-platform audit approach described in the README
- Marks unknown results explicitly instead of guessing when data is missing
- Changes to accounts are draft-first and read-only by default
Trade-offs
- Only Claude Code is verified; other hosts are labeled experimental
- Requires you to supply account evidence yourself, and the quality of audits depends on that data
- Some installs depend on platform-specific requirements such as CPython 3.11 or 3.12 and glibc 2.28 or later on Linux
How it compares
| Project | Stars | Type | Level | Updated | Best for |
|---|---|---|---|---|---|
| claude-ads (this) | 10.7k | AI model / agent | intermediate | 4 days ago | Agencies and consultants reviewing several client ad accounts |
| claude-obsidian | 16.1k | Dev tool | intermediate | 11 days ago | People who already use Obsidian and want an AI assistant to organize incoming notes |
| claude-seo | 25.8k | Dev tool | intermediate | 5 days ago | SEO consultants who want a quick first-call audit with a prioritized fix list |
| career-ops | 79.4k | AI model / agent | intermediate | 4 days ago | Job seekers who apply to many listings and want to filter them before spending time on each one |
| rea | 421k | AI model / agent | intermediate | 4 days ago | Developers studying how a feature in an existing app works without its source code |
| ponytail | 182k | AI model / agent | beginner | 8 days ago | Developers who use Claude Code or similar AI agents and find the generated code too long |
Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.