The Agent Skills Directory

[DATA_EXPOSURE]: The OpenAPI client generation script scripts/generate-openapi.js explicitly sets process.env.NODE_TLS_REJECT_UNAUTHORIZED = '0'. This disables SSL/TLS certificate verification. While common in local Umbraco development to accommodate self-signed certificates, it exposes the user to man-in-the-middle (MITM) attacks if the script is run against a remote or untrusted Swagger endpoint.
[INDIRECT_PROMPT_INJECTION]: The skill contains components that process untrusted data from the external environment.
Ingestion points: notes-browse-dashboard.element.ts extracts folder identifiers directly from the browser's URL path (window.location.pathname). notes-dashboard.element.ts and notes-collection.server.data-source.ts ingest data from API responses.
Boundary markers: Absent. There are no delimiters or instructions to the agent to ignore potentially malicious content inside these data fields.
Capability inventory: The skill possesses capabilities to perform navigation (history.pushState), display notifications, and perform CRUD operations via the NoteswikiService APIs.
Sanitization: Explicit logic-level sanitization is absent; the skill relies on the Lit framework's automatic escaping in HTML templates to prevent XSS.
[DYNAMIC_EXECUTION]: The generate-openapi.js script utilizes the @hey-api/openapi-ts library to generate TypeScript source code at runtime based on an external Swagger definition. This is a form of script generation that incorporates external input into the codebase.
[OBFUSCATION]: Minified JavaScript files are provided in the wwwroot/App_Plugins/ directories (e.g., time-dashboard.js, my-extension.js). These are standard build artifacts for the extensions, and the corresponding source TypeScript code is provided in the Client/src folder.

umbraco-backoffice