最近有哪些 skills
不值得默认信任
这不是热度榜。这里展示的是最近被送来判断、且系统认为需要阻止或至少人工复核的 skills。重点不是它们多流行,而是为什么它们不该被直接装进环境。
fund-daily
Undeclared network API access
research-archive-query
Undeclared subprocess/shell execution
harbor-openclaw
Undeclared network behavior on first load
airoom.ltd-Global-Finance-Data-Platform
HTTP target URL without TLS encryption
face-analysis
Hardcoded Database Credentials in config.yaml
imap-idle-sneder
Hardcoded email credentials in source code
OnionClaw
Missing implementation code—only documentation present
authenticate-wallet
Unversioned npm package execution
heycube-setup
Undeclared persistent hook installation
complianceradar-ai-monitor
Suspicious 'empire-skills' branding
toq
Missing allowed-tools declaration
dygod-movies
Hardcoded NAS credentials in documentation
uplo-legal
Unpinned npm package dependency
boss-ai-assistant
Hardcoded DashScope API Key
LLM Proxy
Critical content-blocking disabled — credential exfiltration not prevented
samantha
Undeclared shell execution via subprocess ping sweep