The Agent Skills Directory

[COMMAND_EXECUTION]: The documentation provides multiple Python scripts using heredoc patterns (python <<'EOF') to interact with the API gateway. These scripts facilitate API operations such as listing audiences, creating connections, and managing subscribers.
[EXTERNAL_DOWNLOADS]: The skill performs network requests to gateway.maton.ai, ctrl.maton.ai, and connect.maton.ai to proxy Mailchimp API calls and manage OAuth authentication flows.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external marketing sources (Mailchimp), creating an attack surface where malicious instructions embedded in subscriber data or campaign content could influence agent behavior.
Ingestion points: API responses from https://gateway.maton.ai/mailchimp/3.0/ endpoints in SKILL.md, specifically subscriber profile fields, campaign HTML content, and plain text body segments.
Boundary markers: Absent; the Python snippets print raw JSON output directly to the agent's context without any delimiters or instructions to treat the data as untrusted.
Capability inventory: The skill allows for comprehensive account management including creating, updating, and deleting audiences, campaigns, and subscribers via urllib.request operations in SKILL.md.
Sanitization: Absent; the provided scripts do not perform any sanitization, filtering, or validation of the fetched content before it is processed by the agent.

mailchimp