Skip to main content
Empower your AI assistants with high-level advertising strategy and operational skills.

​
What are Agent Skills?

Skills are higher-level workflows over raw MCP tools. A tool might perform one API call, such as update_budget. A skill covers a full workflow, such as budget pacing: read data, analyze trends, and propose related actions. ADSMCP workflows start by checking get_connections_status, create campaigns only in paused or approval-required modes, and do not auto-retry failed spend-affecting writes.

Performance Review

Multi-platform health checks that scan for anomalies across connected accounts.

Wasted-spend analysis

Identifies placements, keywords, or creative assets that consume budget without meeting ROAS targets.

Google Search creation

Checks connections, validates the draft, asks for approval, then creates only a paused Search campaign shell.

​
Client support

Codex

Use repo-local plans and skills to enforce connection-first, paused-write workflows.

Claude Code / Cursor

Use copied skill files or client-native skill folders for repeatable advertising workflows.

ChatGPT connector

Connector access can call tools. Workflow enforcement lives in the agent instructions or dashboard playbooks.

​
The brand workspace

Skills rely on local project files to persist context. Keeping these files in your repository keeps the agent aligned with the brand.
  • AGENTS.md defines the persona, tone, and guardrails for the advertising agent.
  • STRATEGY.md stores global KPIs, target CPA, and preferred scaling strategies.