ai-director
Warn
Audited by Socket on Sep 27, 2026
2 alerts found:
Securityx2SecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS: 该技能总体目的与功能基本一致,确实围绕X2C短剧生成展开;但其最敏感的数据流把用户X2C API Key转发到未在文档中充分验证归属的Supabase端点,而非明确官方主域API,这是主要风险。再叠加本地持久化凭证、后台任务跟踪、Telegram/WebChat自动推送等能力,整体风险高于普通文档型技能。未见明确恶意载荷或下载执行链,因此更像高风险/可疑集成而非确认恶意。
Confidence: 85%Severity: 78%
Securityscripts/auto-iterate.js
MEDIUMSecurityMEDIUM
scripts/auto-iterate.js
This is a command-line automation script for generating and evaluating video content. It contains a significant command-injection vulnerability because prompt and style values are interpolated into execSync command strings. The risk is conditional on supplied arguments or evaluator output containing shell syntax; this fragment does not otherwise show clear malicious intent or data theft.
Confidence: 99%Severity: 78%
Audit Metadata