skills/appwrite/skills/appwrite-dart/Gen Agent Trust Hub

appwrite-dart

Pass

Audited by Gen Agent Trust Hub on Oct 23, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill demonstrates patterns for processing untrusted user data which could theoretically be used for indirect prompt injection attacks if not handled carefully by the implementing developer.
  • Ingestion points: Data enters the context via Appwrite Function request bodies (context.req.bodyJson in SKILL.md) and server-side rendering (SSR) login inputs (body['email'] in SKILL.md).
  • Boundary markers: No explicit boundary markers or 'ignore instructions' delimiters are demonstrated in the simplified code examples.
  • Capability inventory: The skill utilizes capabilities including database row creation (tablesDB.createRow), file uploads (storage.createFile), and user authentication (account.createEmailPasswordSession).
  • Sanitization: The provided code examples do not include explicit input validation or sanitization, which represents a potential attack surface though common for documentation examples.
  • [SAFE]: The skill uses official, well-known packages (appwrite, dart_appwrite) and points to the vendor's official domain (appwrite.io).
  • [SAFE]: Sensitive information like API keys is correctly managed using environment variables (Platform.environment['APPWRITE_API_KEY']) rather than being hardcoded.
  • [SAFE]: The documentation includes a dedicated section on 'Permissions & Roles (Critical)' that educates users on avoiding common security pitfalls, such as overly permissive access controls.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 23, 2026, 04:46 AM