firebase-data-connect

Pass

Audited by Gen Agent Trust Hub on Sep 23, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructions utilize npx -y firebase-tools@latest to install and run the Firebase CLI for tasks such as project initialization, emulation, and deployment. This fetches the official tool from the NPM registry, which is a well-known service, and originates from a trusted organization.
  • [COMMAND_EXECUTION]: Development workflows involve executing various shell commands via the Firebase CLI, including firebase init, firebase emulators:start, firebase dataconnect:compile, and firebase deploy. These are standard operations for managing the lifecycle of a Firebase project.
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates the creation of applications that ingest and process user data through GraphQL mutations, representing a surface for indirect prompt injection. To address this, the documentation provides extensive guidance on implementing defensive measures.
  • Ingestion points: External data enters the system through GraphQL mutations defined in connector/mutations.gql and connector/queries.gql filters.
  • Boundary markers: The skill emphasizes the use of @auth directives to define access levels (PUBLIC, USER, NO_ACCESS) and enforce authentication requirements.
  • Capability inventory: The environment allows shell command execution via firebase-tools and file system operations for SDK generation.
  • Sanitization: The skill promotes the use of row-level security through @check expressions using Common Expression Language (CEL) and mandates positional parameters for Native SQL operations to prevent SQL injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 23, 2026, 11:03 PM
Security Audit — agent-trust-hub — firebase-data-connect