secure-python-code-sandbox
Warn
Audited by Socket on Jul 27, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS: the skill’s core capability matches its purpose, and its network endpoints appear to be official AgentPMT infrastructure, so this is not confirmed malware. However, it enables arbitrary remote Python execution, routes user code/data to a hosted third party, and requires transitive installation of other skills, creating meaningful supply-chain and data-handling risk that is higher than a normal documentation-only skill.
Confidence: 86%Severity: 63%
Audit Metadata