subagent-orchestration
Warn
Audited by Socket on May 12, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS: the skill’s purpose matches its orchestration behavior, and it does not show credential theft, exfiltration endpoints, or external download-execute installers. However, it grants the agent broad autonomous control over a codebase after one approval, including file modification, subagent chaining, and execution of repository-defined build/test commands, creating meaningful operational and prompt-injection risk.
Confidence: 100%Severity: 60%
Audit Metadata