antigravity-marketing-lab

Fail

Audited by Gen Agent Trust Hub on Jun 28, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill fetches and executes an installation script from Google's official Antigravity domain (https://antigravity.google/cli/install.sh). This is documented as a standard setup procedure for the documented CLI tool and originates from a trusted organization.
  • [EXTERNAL_DOWNLOADS]: Instructs the user to clone a workshop project from an unverified third-party GitHub repository (github.com/jarsing/antigravity-marketing-lab).
  • [COMMAND_EXECUTION]: Instructions include modifying the user's shell configuration (~/.bashrc) to update the system PATH. While this is standard for CLI tools, it is a persistent modification of the shell environment.
  • [PROMPT_INJECTION]: The skill processes untrusted marketing data from 'challenge/my-brief.md' to generate headlines. This ingestion point lacks explicit boundary markers or sanitization, making the agent susceptible to indirect prompt injection if the user-provided brief contains malicious instructions that override the agent's behavior.
Recommendations
  • HIGH: Downloads and executes remote code from: https://antigravity.google/cli/install.sh - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Jun 28, 2026, 10:52 PM
Security Audit — agent-trust-hub — antigravity-marketing-lab