translation-techniques-specialist

Pass

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No security vulnerabilities or malicious behaviors were detected. The skill consists of instructional content and a workflow for academic research.
  • [PROMPT_INJECTION]: The research checklist in modules/research-checklist.md instructs the agent to ingest external data from sources like Wikipedia and Arxiv. This represents an indirect prompt injection surface; however, it is a documented part of the skill's functionality for maintaining information freshness and targets well-known information repositories.
  • Ingestion points: modules/research-checklist.md (web searches, Wikipedia, Arxiv distillation tools)
  • Boundary markers: None provided in the instruction set
  • Capability inventory: No file system writes, network exfiltration, or command execution capabilities detected
  • Sanitization: Not applicable to the provided static instructions
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 13, 2026, 05:38 PM
Security Audit — agent-trust-hub — translation-techniques-specialist