blackbox
Warn
Audited by Socket on May 16, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS. The skill is broadly aligned with its purpose and uses official Blackbox sources, but it delegates code and prompts to an external AI CLI/service, can ingest untrusted PR content, and documents an auto-approve mode. Main risks are indirect prompt injection and autonomous external-agent behavior rather than confirmed malware.
Confidence: 89%Severity: 62%
Audit Metadata