bm25
Warn
Audited by Socket on Sep 17, 2026
1 alert found:
SecuritySecurityscripts/bm25.py
MEDIUMSecurityMEDIUM
scripts/bm25.py
The code appears to be a legitimate local/GitHub BM25 search utility and contains no clear intentional malware or covert exfiltration. It has material security risks: untrusted pickle deserialization can enable arbitrary code execution if the cache is tampered with, and tar extraction lacks explicit path-traversal protection. Indexing .env and configuration files may expose secrets through normal query output. Use a safe serialization format, validate cache ownership/integrity, and implement secure archive extraction.
Confidence: 97%Severity: 70%
Audit Metadata