k2f-invitation
Warn
Audited by Socket on Oct 1, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
The visible invitation-writing instructions are benign and well-scoped, but the skill is not self-contained: it requires installing a separate third-party skill that will run with the agent's permissions. That transitive trust makes the package suspicious rather than malicious; risk is medium because the main security concern is delegated execution, not direct credential theft or exfiltration in this skill.
Confidence: 93%Severity: 56%
Audit Metadata