managed-agent
Warn
Audited by Socket on May 12, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS. The skill is mostly aligned with its stated purpose of controlling Anthropic Managed Agents, but it enables a powerful remote execution environment with persistent transcripts, arbitrary package/initScript setup, optional third-party MCP connectors, and org-wide session visibility. The main concern is expanded cloud data flow and autonomous remote action capability, not clear malware or deception.
Confidence: 100%Severity: 60%
Audit Metadata