tidas-bilingual-transcreation

Warn

Audited by Snyk on Jul 23, 2026

Risk Level: MEDIUM
Full Analysis

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

  • Third-party content exposure detected (high risk: 0.75). The required workflow ingests outsider-authored free text at runtime from --input /abs/path/rows/processes.jsonl and other “source rows”/evidence fields (outsider dataset content not authored by the operating user), which the LLM reads during the Codex “review” step to generate trans-reviewed.jsonl.

Issues (1)

W011
MEDIUM

Third-party content exposure detected (indirect prompt injection risk).

Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 23, 2026, 08:30 PM
Issues
1
Security Audit — snyk — tidas-bilingual-transcreation