higgsfield-brandkit
Warn
Audited by Socket on Aug 3, 2026
1 alert found:
AnomalyAnomalyscripts/render_brandbook_pdf.py
LOWAnomalyLOW
scripts/render_brandbook_pdf.py
The module itself does not show indicators of intentional malware (no obfuscation, no data exfiltration, no credential theft, no backdoor logic). Its primary security exposure is operational: it executes multiple complex native binaries (fontconfig utilities and LibreOffice) with user-controlled inputs (PPTX path and especially fonts_dir) and inherits the full process environment into those subprocesses. This design is appropriate for a font-constrained conversion pipeline, but it increases risk if an attacker can supply malicious PPTX/font files or influence the runtime environment. A minor exception-handling typo could reduce reliability of error reporting but is not a malicious indicator.
Confidence: 72%Severity: 52%
Audit Metadata