pm-workflow
Warn
Audited by Socket on Mar 24, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS: the core PM coordination behavior is mostly aligned with its purpose, but the skill expands into autonomous downstream orchestration, external research over untrusted sources, and transitive skill execution/modification. There is no clear credential harvesting or malicious exfiltration, yet the combined autonomy and indirect prompt-injection surface make it medium risk.
Confidence: 82%Severity: 60%
Audit Metadata