Skills
skills/akillness/skills-template/firebase-ai-logic/Gen Agent Trust Hub

firebase-ai-logic

Pass

Audited by Gen Agent Trust Hub on Mar 6, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it integrates external data into AI workflows without sufficient isolation.
  • Ingestion points: The analyzeImage function in SKILL.md retrieves content from an external imageUrl.
  • Boundary markers: The skill does not implement delimiters or 'ignore' instructions to separate untrusted data from the system prompt.
  • Capability inventory: The skill possesses network capabilities through the Firebase and Anthropic SDKs to send data to generative models.
  • Sanitization: There is no evidence of input validation or content filtering for fetched external assets.
  • [EXTERNAL_DOWNLOADS]: The skill installs necessary dependencies from trusted organizations.
  • Evidence: It downloads and installs the firebase-tools CLI and the firebase and @anthropic-ai/sdk libraries, which are official packages from Google and Anthropic.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 6, 2026, 12:43 PM