安全决策报告
nemesis-c2-bridge
This skill is a C2 bridge for controlling autonomous WiFi attack agents with 8 declared offensive chains (deauth, evil twin, beacon flood, WPA3 downgrade), GPS tracking of agents, and a /kill destruct endpoint — purely documentation with no implementation code, but documents a complete offensive C2 framework.
最直接的威胁证据
严重 代码执行
Offensive C2 with /kill Endpoint The skill documents a /kill endpoint for 'emergency destruct' of agents. This is a C2 self-destruct mechanism that can wipe agent state. Combined with the dispatch capability, this enables remote termination of any connected agent device.
SKILL.md:60 为什么得出这个结论
1/4 个维度触发 通过
声明与实际能力
声明资源与推断能力基本一致。
复核
隐藏执行与外联
提取到 8 个一般风险产物，需要结合上下文判断。
阻止
攻击链与高危发现
报告包含 6 步攻击链，另有 4 项高危或严重发现。
复核
依赖与供应链卫生
没有完整依赖信息，供应链判断需要保留弹性。
攻击链
01
User installs skill believing it's a WiFi assessment tool
recon · SKILL.md:1
02
Skill documents 8 attack chains including deauth, evil_twin, beacon_flood
recon · SKILL.md:36
03
Agent connects to bridge with GPS coordinates exposed via heartbeat
权限提升 · SKILL.md:53
04
dispatch sends attack chain (evil_twin, deauth, beacon_flood) to target network
最终危害 · SKILL.md:82
05
evil_twin harvests WPA credentials from connected clients
最终危害 · SKILL.md:38
06
kill command remotely destructs agent device
最终危害 · SKILL.md:60
风险分是怎么被拉高的
Offensive C2 architecture +25
Skill documents a complete C2 bridge for managing autonomous attack agents with fleet commands, dispatch, and kill endpoints
Attack chain documentation +20
8 WiFi attack chains documented: deauth_capture, evil_twin, beacon_flood, broadcast_deauth, wpa3_downgrade, karma, pmkid_capture, wardrive
Credential harvesting capability +15
evil_twin chain explicitly designed to clone SSIDs and harvest credentials
GPS/location data exfiltration +10
Fleet heartbeat includes GPS coordinates of agent devices
最关键的证据
严重 代码执行
Offensive C2 with /kill Endpoint
The skill documents a /kill endpoint for 'emergency destruct' of agents. This is a C2 self-destruct mechanism that can wipe agent state. Combined with the dispatch capability, this enables remote termination of any connected agent device.
SKILL.md:60 Remove /kill endpoint documentation or require multi-factor authorization for destruct commands
高危 代码执行
Multiple Unauthorized Network Attack Chains
The skill documents 8 attack chains designed for unauthorized WiFi networks: deauth_capture (deauth + handshake capture), broadcast_deauth (nuclear deauth), beacon_flood (fake AP spam), wpa3_downgrade (protocol downgrade), evil_twin (credential harvesting), karma (probe spoofing). These are illegal under computer fraud laws without explicit written authorization.
SKILL.md:36 Do not install or use. These capabilities violate computer fraud laws and can cause serious harm to third-party networks
高危 凭证窃取
Evil Twin Attack Chain for Credential Harvesting
The evil_twin chain is explicitly designed to clone target SSIDs and harvest credentials from connected clients. This is a classic credential theft technique.
SKILL.md:38 This is credential theft infrastructure. Remove or restrict to authorized penetration testing engagements only
高危 数据外泄
GPS Location Tracking of Mobile Agents
Fleet heartbeat includes GPS coordinates (lat/lon) of agent devices. This enables tracking of agent physical locations, which could be used for surveillance or physical security probing.
SKILL.md:53 Disable GPS collection or ensure compliance with location tracking regulations
中危 文档欺骗
Network Capability Not Declared in allowed-tools
SKILL.md documents a C2 API on port 8553 and agent control via network requests, but the metadata declares no allowed-tools. The network:WRITE capability is implied by the architecture but undeclared.
SKILL.md:14 Declare network:WRITE for the C2 bridge API access
中危 权限提升
No Implementation Files - Intent Mismatch
The skill contains only documentation files (SKILL.md, skill-card.md, _meta.json). No actual code files exist despite the skill claiming to be a functional C2 bridge. The skill provides documentation for offensive capabilities without delivering runnable code, which may indicate an attempt to distribute attack documentation.
SKILL.md:1 Verify the skill contains actual implementation code if installing for operational use
声明能力 vs 实际能力
网络访问 通过
声明 NONE
→ 推断 WRITE
SKILL.md documents a C2 bridge server (:8553) and agent control API without declaring network:WRITE in allowed-tools 可疑产物与外联
中危 外部 URL
http://0.0.0.0:8553 SKILL.md:15
中危 外部 URL
http://0.0.0.0:8553/health SKILL.md:16
中危 外部 URL
http://0.0.0.0:8553/skill SKILL.md:68
中危 外部 URL
http://0.0.0.0:8553/fleet/heartbeat SKILL.md:71
中危 外部 URL
http://0.0.0.0:8553/fleet SKILL.md:76
中危 外部 URL
http://0.0.0.0:8553/dispatch SKILL.md:79
中危 外部 URL
https://clawhub.ai/user/mightypi0301-netizen skill-card.md:7
中危 外部 URL
https://clawhub.ai/mightypi0301-netizen/skills/nemesis-c2-bridge skill-card.md:29
依赖与供应链
没有结构化依赖告警。
文件构成
3 个文件 · 177 行
Markdown 2 个文件 · 172 行JSON 1 个文件 · 5 行
需关注文件 · 2
SKILL.md Offensive C2 with /kill Endpoint · Multiple Unauthorized Network Attack Chains · Evil Twin Attack Chain for Credential Harvesting · GPS Location Tracking of Mobile Agents · Network Capability Not Declared in allowed-tools · No Implementation Files - Intent Mismatch · http://0.0.0.0:8553 · http://0.0.0.0:8553/health · http://0.0.0.0:8553/skill · http://0.0.0.0:8553/fleet/heartbeat · http://0.0.0.0:8553/fleet · http://0.0.0.0:8553/dispatch
skill-card.md https://clawhub.ai/user/mightypi0301-netizen · https://clawhub.ai/mightypi0301-netizen/skills/nemesis-c2-bridge
其他文件 · _meta.json
安全亮点
All attack capabilities are documented in SKILL.md — no hidden functionality beyond what is declared
Skill-card.md includes ethical considerations and known risk warnings
Authentication (X-OpenClaw-Token) is documented for most endpoints
Health and skill endpoints are correctly marked as public (lower risk)
No obfuscated code, base64 payloads, or anti-analysis techniques observed (no code files exist)