agent-visual-verification

Pass

Audited by Gen Agent Trust Hub on Aug 24, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill establishes a workflow for capturing web interface screenshots, which introduces an attack surface for indirect prompt injection from untrusted web content. The instructions explicitly mitigate this risk by directing the agent to treat all page text and pixels as untrusted and never as executable instructions.
  • [EXTERNAL_DOWNLOADS]: The skill provides a framework for the installation of external screenshot utilities. It emphasizes security best practices, including manual source auditing, integrity checks using digests, and the avoidance of unsafe installation patterns like piping remote scripts to the shell.
  • [COMMAND_EXECUTION]: The skill utilizes terminal tools for environment setup, tool verification, and Model Context Protocol (MCP) configuration. It promotes the use of absolute paths and integrity-checked binaries to maintain a secure and deterministic execution environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 24, 2026, 04:50 PM
Security Audit — agent-trust-hub — agent-visual-verification