扫描报告
15 /100
agent-link
智能体互联技能 - 支持不同电脑上的 OpenClaw 实例和 Agent 通过中转服务器进行安全可靠的通讯
A legitimate cross-device agent relay system with no malicious code present; the relay_server.py implements straightforward WebSocket message forwarding with HMAC-SHA256 signing, but documentation references missing files (local-agent component) creating doc-to-implementation gaps.
可以安装
Create the missing scripts/local-agent/ directory with agent_link.py and docs/install-agent.md to match the documented architecture, or update docs to accurately reflect the shipped files.
安全发现 3 项
| 严重性 | 安全发现 | 位置 |
|---|---|---|
| 低危 | Documented local-agent component is missing 文档欺骗 | SKILL.md:72 |
| 低危 | install-agent.md referenced but not present 文档欺骗 | README.md:60 |
| 提示 | websockets dependency not pinned in code 供应链 | scripts/relay-server/relay_server.py:8 |
| 资源类型 | 声明权限 | 推断权限 | 状态 | 证据 |
|---|---|---|---|---|
| 网络访问 | READ | READ | ✓ 一致 | relay_server.py:157 — websockets.serve on 0.0.0.0:port for inbound relay connect… |
| 文件系统 | READ | READ | ✓ 一致 | relay_server.py:193 — reads JSON config file via --config argument |
| 命令执行 | NONE | NONE | — | No subprocess, os.system, or shell execution found in relay_server.py |
| 环境变量 | NONE | NONE | — | No os.environ iteration or environment variable access in the codebase |
| 技能调用 | NONE | NONE | — | No skill invocation or inter-agent call mechanisms present |
目录结构
6 文件 · 21.5 KB · 910 行 Markdown 3f · 634L
Python 1f · 249L
JSON 2f · 27L
├─
▾
docs
│ └─
install-relay.md
Markdown
├─
▾
scripts
│ └─
▾
relay-server
│ ├─
relay_server.py
Python
│ └─
relay-config.example.json
JSON
├─
_meta.json
JSON
├─
README.md
Markdown
└─
SKILL.md
Markdown
依赖分析 1 项
| 包名 | 版本 | 来源 | 已知漏洞 | 备注 |
|---|---|---|---|---|
websockets | * | pip | 否 | Not pinned in requirements file; imported dynamically at runtime |
安全亮点
✓ No base64-encoded execution, eval(), or obfuscation techniques found
✓ No credential harvesting or environment variable enumeration detected
✓ No network exfiltration or C2 communication patterns present
✓ No curl|bash or wget|sh remote script execution patterns
✓ No access to sensitive paths such as ~/.ssh, ~/.aws, or .env files
✓ No persistence mechanisms (cron, startup hooks, systemd units) in code
✓ HMAC-SHA256 signature verification is implemented correctly
✓ Message history deduplication prevents replay attacks
✓ Clean, readable Python implementation with no suspicious code paths
✓ No hardcoded external IPs or suspicious network destinations