think-tank
Warn
Audited by Socket on Aug 13, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS: The skill’s core purpose and capabilities mostly align, but it meaningfully expands the agent’s footprint by orchestrating remote research jobs and consuming their outputs. Main concerns are remote data exposure, environment-dependent broad tool access, and prompt-injection risk from external artifacts/results rather than confirmed malware or deceptive credential theft.
Confidence: 85%Severity: 57%
Audit Metadata