replicate-with-blocks

Pass

Audited by Gen Agent Trust Hub on Jul 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill performs project management and environment setup by executing shell commands and using the WordPress Studio CLI to create sites, manage databases, and activate generated themes.
  • [REMOTE_CODE_EXECUTION]: The skill's primary function is to generate and install executable WordPress theme files (PHP and HTML) based on external website data. This process is protected by a mandatory validation gate (liberate_validate_artifacts) that scans for and blocks unauthorized PHP tags, script tags, and event handlers to prevent code injection.
  • [EXTERNAL_DOWNLOADS]: The skill downloads assets such as font files and images discovered during the site extraction process to ensure the final theme is self-hosted and does not rely on third-party content delivery networks.
  • [PROMPT_INJECTION]: While processing untrusted HTML and content from external sites, the skill is exposed to indirect prompt injection. It mitigates this risk by enforcing a provenance check during validation, ensuring the agent only uses captured data for its intended structural purpose and does not obey embedded instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 26, 2026, 03:02 AM
Security Audit — agent-trust-hub — replicate-with-blocks