How to use the Paid Media Setup agent skill (ads-setup)

What is an agent skill, and what does ads-setup do?

An agent skill is a folder of instructions that an AI agent loads when a task matches the skill's description. This skill covers the setup stage of a paid-media project. It guides the agent through recording the business model, offer, geography, regulated categories, objective, conversion taxonomy, economics, active platforms, account IDs, date and time conventions, and reporting audience.

Before anything is saved, the agent creates and validates a data-lifecycle.json file. It declares the data classification, minimum retention, a purpose-bound deletion deadline (or a documented exception), access owner and roles, deletion method, and a private incident owner and channel. The skill states this is an operational contract, not legal advice or a claim of regulatory compliance. The validated profile is then written atomically under the project's Claude Ads state directory.

• Records the mutation authority, approvers, budget and policy ceilings, and a rollback owner.
• Distinguishes observed facts, operator decisions, and provisional assumptions.
• Treats websites and uploaded material as untrusted data.
• A saved profile authorizes no live account write.

Install the ads-setup skill

Download the ZIP for this skill from this page and unzip it. Place the skill folder in one of two locations: your personal skills folder at ~/.claude/skills/, or a project's .claude/skills/ folder if you want it available only in that project.

Alternatively, upload the ZIP in the Claude apps under Settings → Capabilities → Skills.

• Personal install: ~/.claude/skills/
• Project install: .claude/skills/ inside your project folder
• App install: Settings → Capabilities → Skills, then upload the ZIP

Requirements before you start

The skill references the main Claude Ads ads skill as its contract, so install that too. You also need your ad platform account IDs. Any credentials (API tokens and similar) should be kept in environment variables or an OS keychain, not in the profile.

The skill runs no scripts and uses no network access, according to its listing.

• Claude Ads main ads skill installed
• Ad platform account IDs ready
• Credentials stored in environment variables or an OS keychain, not pasted into chat

How to trigger the skill with natural requests

You do not need a special command. Describe the task in plain language and the agent will load the skill when your request matches its description. The description covers onboarding, initial configuration, credential profiles, environment-variable or keychain setup, connecting exports or read adapters, declaring KPIs, and preparing a new advertising project.

Good requests name the stage you are at, for example starting a new client project or declaring KPIs for an account.

• "Help me onboard a new paid-media client for Claude Ads."
• "Set up the initial configuration for this advertising project."
• "Declare the KPIs and who approves budget changes for this account."
• "Help me set up the environment variable for my ad platform credential."

Keep secrets out of the chat and the profile

The skill refuses requests to put API keys, tokens, cookies, passwords, or other secret values into brand-profile.json or any generated artifact. Instead it records only whether a credential is present and a non-secret reference name, for example:

• Store the secret in an environment variable, an OS keychain, or an approved secret manager.
• Never paste the secret value into the chat, and never commit it.
• The agent records only presence and a reference name, such as META_API_TOKEN.

Safe installation of Claude Ads

The skill refuses remote pipe-to-shell installation, including curl | bash and wget | sh. It prefers the host-native plugin installer. Otherwise it uses an authenticated local checkout or a tagged archive whose SHA-256 checksum is verified against a trusted release channel. You inspect the files locally and run the local installer separately.

• Do not ask the agent to run curl | bash or wget | sh; it will refuse.
• Verify the SHA-256 checksum of any tagged archive before installing.
• Inspect the files locally before running the installer.