Skills
skills/inferencesh/skills/google-veo/Gen Agent Trust Hub

google-veo

Fail

Audited by Gen Agent Trust Hub on Feb 18, 2026

Risk Level: CRITICALREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • REMOTE_CODE_EXECUTION (CRITICAL): The skill documentation explicitly instructs the agent to run curl -fsSL https://cli.inference.sh | sh. This pattern downloads and executes a script from an untrusted domain without validation, allowing for arbitrary code execution.
  • EXTERNAL_DOWNLOADS (HIGH): The skill relies on external tools and additional skill sets hosted on inference.sh, which is not a trusted source according to the safety guidelines.
  • COMMAND_EXECUTION (MEDIUM): The skill uses the Bash tool to execute infsh commands. This permission allows the agent to interact with the untrusted binary downloaded during the setup phase.
  • PROMPT_INJECTION (LOW): This finding identifies an indirect prompt injection surface. 1. Ingestion points: User-provided prompts in infsh app run commands within SKILL.md. 2. Boundary markers: JSON formatting is used but no 'ignore embedded instructions' markers are present. 3. Capability inventory: Subprocess execution via Bash. 4. Sanitization: No input sanitization detected.
Recommendations
  • HIGH: Downloads and executes remote code from: https://cli.inference.sh - DO NOT USE without thorough review
  • AI detected serious security threats
Audit Metadata
Risk Level
CRITICAL
Analyzed
Feb 18, 2026, 11:14 AM