The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the official Mayar CLI via npx -y mayar@latest. This allows for account management tasks such as checking balances, creating invoices, and managing webhooks directly from the terminal.
Evidence: Found in SKILL.md and references/cli-commands.md.
[EXTERNAL_DOWNLOADS]: The skill dynamically fetches documentation and API schema information from https://docs.mayar.id/llms.txt and specific endpoint markdown files on the same domain to ensure the integration uses up-to-date API facts.
Evidence: Found in references/api-sources.md and references/product-knowledge.md.
[INDIRECT_PROMPT_INJECTION]: The skill defines an attack surface by ingesting external data from the Mayar documentation site to guide the agent's behavior during the planning and implementation phases.
Ingestion points: Reads content from https://docs.mayar.id/llms.txt and associated .md pages (specified in references/api-sources.md).
Boundary markers: The skill uses a structured phase-gate process (Discover, Plan, Implement, Verify) to control data flow, though explicit 'ignore instructions' delimiters for the fetched content are absent.
Capability inventory: The agent can execute shell commands via npx, write files to the local project, and create network requests to the Mayar API (documented in playbook/implement.md and references/cli-commands.md).
Sanitization: The skill explicitly instructs the developer to implement input validation and error handling in the generated code (e.g., validateCheckoutRequest in references/stack-nextjs.md) and emphasizes the use of official schema facts only.

mayar-v2