boxlang-modules-and-packages
Pass
Audited by Gen Agent Trust Hub on Oct 20, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill describes installing modules via the CommandBox
boxCLI from the official ForgeBox registry, which is standard ecosystem behavior. - [CREDENTIALS_UNSAFE]: Configuration examples demonstrate the use of environment variable placeholders (e.g.,
${REDIS_PASSWORD},${SMTP_PASS}) instead of hardcoded credentials, following security best practices. - [INDIRECT_PROMPT_INJECTION]: The skill documents capabilities for ingesting external data, which represents a potential attack surface for indirect prompt injection. 1. Ingestion points:
csvParse,spreadsheetRead,bx:ldap, andimageReadfunctions inSKILL.mdprocess external files or directory services. 2. Boundary markers: The documentation examples do not define boundary markers for untrusted data. 3. Capability inventory: The skill documents capabilities including file system writes (fileWrite,spreadsheetWrite,imageWrite), PDF generation (bx:pdf), and network operations (bx:mail,bx:ldap,RedisService). 4. Sanitization: No specific sanitization or escaping logic is shown in the instructional snippets. - [SAFE]: The skill is instructional and its behavior aligns with its primary purpose of documenting official vendor framework capabilities.
Audit Metadata