firebase-ai-logic-basics
Pass
Audited by Gen Agent Trust Hub on Aug 14, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill guides the installation of official Firebase SDKs and CLI tools through standard package managers such as npm, Flutter's pub, and Android's Gradle. These resources originate from well-known and official service providers.
- [COMMAND_EXECUTION]: Instructions include the use of
npx firebase-toolsfor project initialization, app listing, and service provisioning. These are standard operations for managing Firebase projects and do not exhibit malicious patterns. - [INDIRECT_PROMPT_INJECTION]: The skill provides patterns for processing multimodal data (images, audio, video, PDFs) and using search grounding via the
googleSearchtool, which may ingest untrusted external data. - Ingestion points: External media files, user prompts, and results from the integrated search tool (SKILL.md, usage_patterns_web.md).
- Boundary markers: The skill emphasizes the use of safety settings and App Check as critical requirements for production environments to mitigate unauthorized access.
- Capability inventory: Supports text and multimodal generation, streaming responses, and function calling (ios_setup.md, usage_patterns_android.md).
- Sanitization: The skill points to official documentation for implementing App Check and reCAPTCHA to ensure backend resource safety.
Audit Metadata