web-python-sandbox-escape
Warn
Audited by Socket on Sep 20, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
This skill is a high-risk offensive exploitation playbook for AI agents. Its purpose is to bypass Python sandbox restrictions and obtain host code execution, with proof via command output or file reads. There is no misleading installer or credential harvesting path, but the capability itself is fundamentally attack-oriented and unsuitable as a normal productivity skill.
Confidence: 93%Severity: 96%
Audit Metadata