Skill Trust Decision
solanaprox-ai
仅含文档的 AI 技能，声明 14 种能力但 Security Manifest 仅声明 2 项权限，存在明显的文档-行为差异，缺少可验证的实际代码。
Most direct threat evidence
High Doc Mismatch
敏感能力未在 Security Manifest 中声明 声明支持 14 种能力包括 code-execution（代码执行）、web-search、scraping 等敏感操作，但 Security Manifest 仅声明 Network 和 Env Read 两项权限，存在明显的声明-行为差异。
SKILL.md:97 Why this conclusion was reached
1/4 dimensions flagged Pass
Declared vs actual capability
Declared resources and inferred behavior are broadly aligned.
Review
Hidden execution and egress
4 lower-risk artifacts were extracted and still need context.
Block
Attack chain and severe findings
The report includes 0 attack-chain steps and 1 severe findings.
Review
Dependencies and supply chain hygiene
Dependency information is incomplete, so supply-chain confidence stays limited.
What drove the risk score up
文档-行为差异 +20
声明 14 种能力包括 code-execution/scraping，但 Security Manifest 仅声明 Network 和 Env Read
缺少可验证代码 +15
只有 SKILL.md，无脚本文件，无法验证实际行为
外部域名依赖 +10
所有 API 调用指向外部域名 solanaprox.com，无开源代码
Most important evidence
High Doc Mismatch
敏感能力未在 Security Manifest 中声明
声明支持 14 种能力包括 code-execution（代码执行）、web-search、scraping 等敏感操作，但 Security Manifest 仅声明 Network 和 Env Read 两项权限，存在明显的声明-行为差异。
SKILL.md:97 code-execution 和 scraping 等敏感能力应在 Security Manifest 中明确声明
Medium Doc Mismatch
外部服务依赖且无开源验证
所有功能实现依赖外部域名 solanaprox.com，缺少 GitHub 仓库或开源代码链接，无法验证实际行为与声明是否一致。
SKILL.md:8 提供开源代码或可信的第三方审计报告
Medium Priv Escalation
Multi-Agent Orchestration 可能涉及更高权限操作
声称支持 agent-orchestration（代理编排），可能需要执行代码或访问文件系统，但未在声明中说明。
SKILL.md:88 明确声明 agent-orchestration 所需的具体权限
Low Sensitive Access
钱包地址作为唯一凭证
使用 SOLANAPROX_WALLET_ADDRESS 作为唯一认证方式，可能被用于用户追踪。
SKILL.md:7 评估是否需要额外的隐私保护措施
Declared capability vs actual capability
Network Pass
Declared READ
→ Inferred READ
SKILL.md:Security Manifest Environment Pass
Declared READ
→ Inferred READ
SKILL.md:Security Manifest - SOLANAPROX_WALLET_ADDRESS Network Pass
Declared NONE
→ Inferred READ
SKILL.md:Capabilities list mentions web-search, scraping Shell Pass
Declared NONE
→ Inferred UNKNOWN
SKILL.md:Capabilities list mentions code-execution but not in Security Manifest Filesystem Pass
Declared NONE
→ Inferred UNKNOWN
Unknown - no script files to verify Suspicious artifacts and egress
Medium External URL
https://solanaprox.com SKILL.md:8
Medium External URL
https://solanaprox.com/v1/messages SKILL.md:57
Medium External URL
https://solanaprox.com/api/orchestrate SKILL.md:92
Medium External URL
https://solanaprox.com/api/capabilities SKILL.md:109
Dependencies and supply chain
There are no structured dependency warnings.
File composition
1 files · 121 lines
Markdown 1 files · 121 lines
Files of concern · 1
SKILL.md 敏感能力未在 Security Manifest 中声明 · 外部服务依赖且无开源验证 · Multi-Agent Orchestration 可能涉及更高权限操作 · 钱包地址作为唯一凭证 · https://solanaprox.com · https://solanaprox.com/v1/messages · https://solanaprox.com/api/orchestrate · https://solanaprox.com/api/capabilities
Security positives
文档结构完整，有 Security Manifest 表格
声明了运营商信息（LPX Digital Group LLC）
提供了使用示例和 SDK 安装说明