expert-douyin
Warn
Audited by Socket on Oct 21, 2026
3 alerts found:
Securityx3Securitytools/douyin-publish/SKILL.md
MEDIUMSecurityMEDIUM
tools/douyin-publish/SKILL.md
结论为 SUSPICIOUS。技能目的与能力基本一致:本地浏览器自动化发布抖音内容、保留会话、短信验证续接、结果核查都符合“发布工具”定位,且未发现明显把凭据转发到无关第三方端点的说明。主要风险在于技能实际依赖一个来源未验证的 `douyin-publish` 本地 CLI;它会接触本地文件、持久化会话和真实发布动作,但技能未提供其官方发行证明或安装链,因此属于高供应链风险而非已确认恶意。
Confidence: 89%Severity: 78%
SecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS. The stated purpose matches Douyin operations, but the skill’s core execution depends on unverifiable custom CLIs and forwards high-value session material into them. Camoufox itself appears legitimate, yet the undocumented wrappers and opaque auth/data endpoints create a high supply-chain and credential-handling risk.
Confidence: 79%Severity: 84%
Securitytools/douyin-engagement/SKILL.md
MEDIUMSecurityMEDIUM
tools/douyin-engagement/SKILL.md
SUSPICIOUS:技能目的与“已发布作品取数”总体一致,但其关键实现依赖读取本地浏览器登录态,并将补数流程建立在无法确认官方归属的 OFB Relay 签名链路上;同时核心 CLI/写库工具缺少充分可验证的发布来源。未见明确恶意或隐蔽窃密指令,但供应链与凭据流风险偏高。
Confidence: 82%Severity: 74%
Audit Metadata