dspy-ruby

Fail

Audited by Gen Agent Trust Hub on Feb 15, 2026

Risk Level: HIGHPROMPT_INJECTION
Full Analysis
  • PROMPT_INJECTION (HIGH): The templates define significant Indirect Prompt Injection surfaces where untrusted external data is ingested and processed by LLMs with tool-execution capabilities.
  • Ingestion points: assets/module-template.rb (BasicClassifier, ResearchAgent) and assets/signature-template.rb (DocumentParser).
  • Boundary markers: No delimiters or isolation markers are present in the signatures or module templates to prevent untrusted data from overriding instructions.
  • Capability inventory: ResearchAgent utilizes DSPy::ReAct for tool invocation (SearchTool, FinishTool), and templates include configurations for model selection and logging.
  • Sanitization: No input sanitization or validation is implemented in the provided boilerplate.
  • Automated Scan False Positive (LOW): The scanner report flagging Rails.logger.info as a malicious URL is incorrect; this is a standard Ruby on Rails method for logging and does not represent a network threat.
Recommendations
  • AI detected serious security threats
  • Contains 1 malicious URL(s) - DO NOT USE
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 15, 2026, 10:16 PM