bx-sites-plugins

Pass

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill describes using the bxSites install:plugin command, which downloads and installs BoxLang modules from the ForgeBox package registry. ForgeBox is the official and well-known registry for the BoxLang and Ortus Solutions ecosystem.
  • [DYNAMIC_EXECUTION]: The plugin architecture involves dynamically loading BoxLang classes (e.g., BxSitesPlugin.bx and BxSitesCliProvider.bx) into the runtime environment. These modules execute custom logic during the site build process (hooks like onConfig, onPageHtml, etc.) or when specific CLI verbs are called.
  • [COMMAND_EXECUTION]: The documentation provides examples for executing CLI commands such as bxSites install:plugin and bxSites plugin:new. These commands are used for managing the development and installation of extension modules.
  • [INDIRECT_PROMPT_INJECTION]: The skill documents hooks that process content from external files, which represents a vulnerability surface where instructions embedded in data could affect agent operations.
  • Ingestion points: onPageMarkdown and onPageHtml methods in models/BxSitesPlugin.bx receive raw content from project documentation files.
  • Boundary markers: None specified; hooks operate directly on the provided string data.
  • Capability inventory: Plugins can perform file system operations, manipulate configuration structures, and execute arbitrary BoxLang code, including network calls if implemented within the module.
  • Sanitization: The reference documentation does not specify sanitization requirements; security is dependent on the implementation of the individual plugin.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 28, 2026, 03:14 AM
Security Audit — agent-trust-hub — bx-sites-plugins