pixelbrowse

Warn

Audited by Gen Agent Trust Hub on Oct 18, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to install an external package named pixelrag from a public package registry using pip, uv, or pipx. This package is not provided by a known trusted organization, and its source code is not verified within the skill.
  • [COMMAND_EXECUTION]: The skill provides instructions for executing various shell commands using pixelshot and python3, which include network-connected operations and file system manipulations.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to capture and visually process content from arbitrary URLs, creating a vulnerability surface where an attacker-controlled website could present instructions to the agent.
  • Ingestion points: The pixelshot <url> command captures screenshots from remote, potentially untrusted websites (SKILL.md).
  • Boundary markers: The instructions do not specify any delimiters or warnings to ignore instructions found within the visual content of the captured pages.
  • Capability inventory: The skill utilizes the Bash tool for command execution and the Read tool for file access.
  • Sanitization: There is no mention of sanitization, filtering, or validation for the content rendered from the URLs before the agent processes it.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Oct 18, 2026, 10:31 AM