active-storage

Pass

Audited by Gen Agent Trust Hub on May 1, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides examples for attaching remote files using URI.open. While this is a standard feature of Rails Active Storage, it is noted as a potential surface for Server-Side Request Forgery (SSRF) if implemented with unsanitized user-provided URLs.
  • [COMMAND_EXECUTION]: The instructions include the use of bin/rails runner for configuration checks and system for executing ffprobe to extract video metadata. These commands are typical for Rails application management and asset processing.
  • [PROMPT_INJECTION]: The skill handles untrusted data through file uploads, establishing an indirect prompt injection surface.
  • Ingestion points: Model attachments via params.expect in SKILL.md (Step 4).
  • Boundary markers: No specific markers or delimiters are defined for the binary content of the attachments.
  • Capability inventory: Includes filesystem access, database migrations, and subprocess calls for media analysis (e.g., ffprobe, libvips).
  • Sanitization: The skill explicitly recommends implementing content type and file size validations in the model as a mitigation strategy.
Audit Metadata
Risk Level
SAFE
Analyzed
May 1, 2026, 04:09 AM