review-ugc-render
Warn
Audited by Socket on Jul 15, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS. The core capability is well aligned with the stated QC purpose, and the local tooling footprint is proportionate. The main risk is data-flow integrity: honoring OPENAI_BASE_URL and naming an unverifiable Whisper proxy means audio and API credentials may be routed to a third-party endpoint instead of official OpenAI services.
Confidence: 89%Severity: 56%
Audit Metadata