The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The file references/litestar_integration.md includes a hardcoded secret development-only-secret-32-chars used in a CookieBackendConfig example. While explicitly labeled for development, its presence in the documentation could lead to insecure configurations if used in production.
[COMMAND_EXECUTION]: The InertiaSSRConfig documented in both SKILL.md and references/litestar_integration.md allows the definition of a command, such as ["node", "resources/ssr.js"], which the plugin executes as a subprocess to handle server-side rendering.
[EXTERNAL_DOWNLOADS]: The skill references the litestar assets install CLI command, which typically fetches external packages from public registries like NPM.
[DYNAMIC_EXECUTION]: The integration supports runtime execution of JavaScript files for SSR via a Node.js process as part of its core functionality.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates the transfer of data from Python route handlers to client-side JavaScript components, creating a surface where malicious server-side data could influence frontend behavior.
Ingestion points: Data returned by Python route handlers (e.g., DashboardController.index) and msgspec DTOs.
Boundary markers: Data is passed via the Inertia JSON protocol; there are no specific boundary markers or instructions to ignore embedded content within the shared props.
Capability inventory: Command execution capabilities via SSR processes and asset management CLI tools.
Sanitization: The skill relies on the automatic output escaping and sanitization provided by the React, Vue, or Svelte client-side adapters.

litestar-inertia