The Agent Skills Directory

[SAFE]: The skill demonstrates standard security practices by recommending the use of environment variables for API keys and specifying secure attributes (HttpOnly, Secure, SameSite=Strict) for session cookies.- [PROMPT_INJECTION]: Indirect Prompt Injection Surface. The documentation defines patterns for ingesting and processing untrusted external data, establishing an attack surface. 1. Ingestion points: Data enters the agent context through context.req.bodyJson and request.uri.queryParameters in server-side logic examples. 2. Boundary markers: No explicit boundary markers or delimiters are suggested in the provided code templates. 3. Capability inventory: The skill provides extensive capabilities for database management, user administration, and file storage. 4. Sanitization: The documentation focuses on secure transport and storage configuration rather than explicit input validation or escaping.- [SAFE]: Deceptive Technical Information. The skill contains misleading metadata by claiming the standard 'Databases' class is deprecated and recommending a non-existent 'TablesDB' class. This factually incorrect information for the Appwrite SDK could lead to development errors or the use of non-standard, potentially unsafe resources.

appwrite-dart