The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a standalone Node.js script at assets/scripts/setup-access-app.ts. This script is intended to be executed by the agent to interact with the Cloudflare API for provisioning Access applications and policies. It contains logic to verify existing configurations before making updates, ensuring idempotent operations.
[EXTERNAL_DOWNLOADS]: The documentation provides instructions to install the skill from the author's public GitHub repository (mizchi/skills) using tools like apm or npx skills. These methods are standard for distributing and managing AI agent skills.
[INDIRECT_PROMPT_INJECTION]: The script contains an ingestion surface for external data via environment variables like APP_ACCESS_ALLOWED_EMAILS (found in assets/scripts/setup-access-app.ts).
Ingestion points: Environment variables provided to the script at runtime.
Boundary markers: The ingested data is strictly used as value fields within a JSON payload for the Cloudflare API and is not processed as instructions for the agent.
Capability inventory: The script is capable of performing network requests to Cloudflare's official API endpoints (api.cloudflare.com) using the global fetch API.
Sanitization: The script includes a validateEmails function that checks the format of the ingested email addresses against a regular expression before they are included in any API requests.

cloudflare-access-app-setup