low-altitude-guardian

低空无人设备应急裁决引擎。零依赖可用:基于损失优先级金字塔(P0-P4)和加权决策公式,对无人机/eVTOL突发危机进行分级分析、方案推导、输出可执行决策建议。分析辅助工具,不连接飞控系统,不执行实际飞行控制。

3,891 stars

Best use case

low-altitude-guardian is best used when you need a repeatable AI agent workflow instead of a one-off prompt.

低空无人设备应急裁决引擎。零依赖可用:基于损失优先级金字塔(P0-P4)和加权决策公式,对无人机/eVTOL突发危机进行分级分析、方案推导、输出可执行决策建议。分析辅助工具,不连接飞控系统,不执行实际飞行控制。

Teams using low-altitude-guardian should expect a more consistent output, faster repeated execution, less prompt rewriting.

When to use this skill

  • You want a reusable workflow that can be run more than once with consistent structure.

When not to use this skill

  • You only need a quick one-off answer and do not need a reusable workflow.
  • You cannot install or maintain the underlying files, dependencies, or repository context.

Installation

Claude Code / Cursor / Codex

$curl -o ~/.claude/skills/low-altitude-guardian/SKILL.md --create-dirs "https://raw.githubusercontent.com/openclaw/skills/main/skills/aaalenwow/low-altitude-guardian/SKILL.md"

Manual Installation

  1. Download SKILL.md from GitHub
  2. Place it in .claude/skills/low-altitude-guardian/SKILL.md inside your project
  3. Restart your AI agent — it will auto-discover the skill

How low-altitude-guardian Compares

Feature / Agentlow-altitude-guardianStandard Approach
Platform SupportNot specifiedLimited / Varies
Context Awareness High Baseline
Installation ComplexityUnknownN/A

Frequently Asked Questions

What does this skill do?

低空无人设备应急裁决引擎。零依赖可用:基于损失优先级金字塔(P0-P4)和加权决策公式,对无人机/eVTOL突发危机进行分级分析、方案推导、输出可执行决策建议。分析辅助工具,不连接飞控系统,不执行实际飞行控制。

Where can I find the source code?

You can find the source code on GitHub using the link provided at the top of the page.

Related Guides

SKILL.md Source

接收突发情况描述,输出结构化应急决策建议。

**⚠️ ALPHA — 重要声明:**
- 本技能为**分析辅助工具**,不连接任何飞控系统,不执行实际飞行控制
- 输出结果为决策参考建议,不替代经认证的飞行安全系统
- 真实飞行作业中,任何最终决策须由持证操作员或认证飞控系统执行

---

## 运行模式

| 模式 | 依赖 | 适用场景 |
|------|------|---------|
| **推理分析模式**(默认) | 无 | 情景演练、方案推导、培训、Agent 集成 |
| **集成模式** | Python + 数据接口 | 与实际遥测系统对接(开发/仿真环境) |

---

## 调用示例

```
# 单次危机分析
我的无人机 GPS 信号突然丢失,当前高度 80m,电量 45%,下方是居民区。帮我分析应对方案。

# 结构化输入
设备: UAV-001 多旋翼
位置: 高度120m,速度15m/s,下方有人行道
状态: 电量34%,左前电机异常振动
环境: 风速8m/s,通信正常
触发: 电机异常,机身开始偏航

# 方案推演
如果同时发生 GPS 丢失 + 单电机故障,应该怎么处置?

# 知识库查询
电池热失控的标准处置流程是什么?
```

---

## 核心方法论:损失优先级金字塔

**所有决策的唯一不可动摇约束:**

```
P0 人员安全   ← 绝对最高优先级,任何情况下不可妥协
P1 公共安全   ← 公共设施、建筑、交通
P2 第三方财产 ← 他人车辆、农作物
P3 本机安全   ← 设备自身完整性
P4 任务完成   ← 原定任务目标
```

**关键原则**:为保障 P0,可以主动坠毁设备(P3)。高优先级受威胁时,必须无条件牺牲低优先级。

---

## Phase 1: 态势快照

接收突发情况后,立即提取以下关键字段(自然语言描述或结构化均可):

```
设备: [device_id] [device_type]
位置: [坐标/区域描述] 高度[m] 速度[m/s] 航向
状态: 电量[%] 飞行阶段 载荷[kg]
触发: [危机触发原因]
环境: 风速[m/s] 下方区域[人群/建筑/空旷] 通信[正常/异常]
```

同步推断:可飞行安全包络线、最近可用备降点、周边禁飞区。

---

## Phase 2: 危机分级

| 等级 | 名称 | 判定条件 | 分析时限 |
|------|------|---------|---------|
| **L5-CRITICAL** | 灾难性 | 即刻威胁人员生命 | **< 3秒** |
| **L4-SEVERE** | 严重 | 高概率人员伤害或重大财产损失 | < 10秒 |
| **L3-MAJOR** | 重大 | 设备功能严重降级,可能危及安全 | < 30秒 |
| **L2-MINOR** | 一般 | 功能部分降级,可控范围 | < 2分钟 |
| **L1-CAUTION** | 注意 | 潜在风险,暂无直接威胁 | < 5分钟 |

**态势演化预测**:预测 30s/60s/180s 后态势,识别级联故障风险(如电机故障 → 电池过放)和可用决策时间窗口。

---

## Phase 3: 方案推导与最优裁决

### 3.1 知识库匹配

覆盖场景:动力系统故障 / 导航定位故障 / 通信故障 / 环境威胁 / 碰撞风险 / 复合故障

```bash
# 集成模式可调用:
python3 scripts/decision_manager.py --match --crisis-type <type> --level <level>
```

### 3.2 最优裁决公式

```
Score = 0.40×S₀(人员安全) + 0.25×S₁(公共安全) + 0.15×S₂(第三方财产)
      + 0.12×S₃(本机安全) + 0.08×S₄(任务完成)

硬约束:S₀ < 80 的方案一律淘汰,无论总分多高
```

### 3.3 无匹配时的第一性原理推理

新型故障或罕见组合时:

1. **分解** → 拆解为已知子问题组合
2. **评估可控能力** → 当前设备还有哪些控制能力
3. **最低降落需求** → 安全着陆至少需要什么能力
4. **无法安全降落时** → 按 P0→P1→P2→P3 顺序选择损失最小的坠落区域
5. **保守原则** → 无先例时,选最保守方案

### 3.4 人工介入模式参考

| 等级 | 建议介入模式 |
|------|------------|
| L5 | 自主执行,事后通知(无等待时间) |
| L4 | 自主执行,同步通知操作员 |
| L3 | 推荐方案,等待 5 秒确认,超时自动执行 |
| L2 | 推荐方案,等待操作员确认 |
| L1 | 仅提醒,由操作员决策 |

---

## 输出格式

```json
{
  "crisis_level": "L3-MAJOR",
  "crisis_type": "power_failure.single_motor_loss",
  "time_window_seconds": 30,
  "selected_plan": {
    "name": "三电机降级返航",
    "source": "knowledge_base | first_principles",
    "confidence": 0.87,
    "steps": [
      {"seq": 1, "action": "切换三电机飞行模式", "timeout_ms": 500},
      {"seq": 2, "action": "降至安全包络高度", "timeout_ms": 5000},
      {"seq": 3, "action": "计算最近备降点航线", "timeout_ms": 1000},
      {"seq": 4, "action": "降级模式飞往备降点", "timeout_ms": null},
      {"seq": 5, "action": "执行应急着陆程序", "timeout_ms": 30000}
    ],
    "priority_score": 91.2,
    "estimated_loss": {
      "human_safety_risk": "none",
      "public_safety_risk": "none",
      "device_damage": "minor"
    }
  },
  "eliminated_plans": [
    {"name": "强制立即降落", "reason": "S₀=65 < 80,下方有行人,一律淘汰"}
  ]
}
```

---

## Phase 4: 事件复盘与知识库自迭代

事件结束后:

```bash
python3 scripts/incident_reporter.py --generate-report --incident-id <id>
python3 scripts/knowledge_updater.py --learn --incident-id <id>
```

**复盘内容**:触发原因、决策过程(选了什么/为什么/排除了什么)、执行效果、实际损失。

**知识库更新**:第一性原理推导出的有效方案,自动封装为新模板。每次事件都让系统更强。

---

## 边界与合规

**本技能做什么:** 危机情景分析、方案推导、决策评分、复盘报告。

**本技能不做什么:**
- 不连接任何飞控系统或硬件接口
- 不执行实际飞行控制指令
- 不替代经 CAAC/FAA 认证的飞行安全系统
- 不提供法律或保险合规建议

**数据流向**:所有情景分析在本次会话中处理,不向外部系统传输飞行数据。集成模式下的遥测数据处理范围由用户自行配置。

Related Skills

security-guardian

3891
from openclaw/skills

Automated security auditing for OpenClaw projects. Scans for hardcoded secrets (API keys, tokens) and container vulnerabilities (CVEs) using Trivy. Provides structured reports to help maintain a clean and secure codebase.

Security

guardian-wall

3891
from openclaw/skills

Mitigate prompt injection attacks, especially indirect ones from external web content or files. Use this skill when processing untrusted text from the internet, user-uploaded files, or any external source to sanitize content and detect malicious instructions (e.g., "ignore previous instructions", "system override").

Security

session-guardian

3891
from openclaw/skills

Never lose a conversation again. Auto-backup, smart recovery, and health monitoring for OpenClaw sessions. Protects against gateway crashes, model disconnections, and token overflow. Use this skill when: - User worries about losing conversations after gateway restart or model crash - User mentions session backup, conversation recovery, session protection, or data loss - User's agent is slow or timing out (likely token overflow from large sessions) - User runs multiple agents and needs to track collaboration across sessions - User asks about session health, backup strategy, or disaster recovery - User mentions "对话丢失", "会话备份", "上下文溢出", "token超限", "Gateway重启后记忆丢失" - Even if user just says "my agent lost everything after a restart" — this is the skill

General Utilities

solidity-guardian

3891
from openclaw/skills

Smart contract security analysis skill. Detect vulnerabilities, suggest fixes, generate audit reports. Supports Hardhat/Foundry projects. Uses pattern matching + best practices from Trail of Bits, OpenZeppelin, and Consensys.

agency-guardian

3891
from openclaw/skills

Gentle reminders to stay human while using AI. Reflection, not restriction.

config-guardian

3891
from openclaw/skills

Safe OpenClaw config updates with automatic backup, validation, and rollback. For agent use - prevents invalid config updates.

personal-guardian

3891
from openclaw/skills

个体终端应急守护智能体(PTERA)。当用户激活"安全时刻"或设备自动检测到危险信号时,Agent 获得完全自主决策权,默认用户处于无法应答状态,以人身安全为最高优先级执行饱和式救援——录音、定位、联系人链式通知、120/110 自主呼叫、无人机急救网络联动。

reminder-guardian

3891
from openclaw/skills

Helps you remember things by keeping a list of reminders, creating the scheduled jobs to alert you, and tracking which ones are done.

cron-job-guardian

3891
from openclaw/skills

检查 cron 或 timer 配置中的频率、幂等、重试、日志与并发风险。;use for cron, timer, ops workflows;do not use for 直接启停生产任务, 替代真正监控.

---

3891
from openclaw/skills

name: article-factory-wechat

Content & Documentation

humanizer

3891
from openclaw/skills

Remove signs of AI-generated writing from text. Use when editing or reviewing text to make it sound more natural and human-written. Based on Wikipedia's comprehensive "Signs of AI writing" guide. Detects and fixes patterns including: inflated symbolism, promotional language, superficial -ing analyses, vague attributions, em dash overuse, rule of three, AI vocabulary words, negative parallelisms, and excessive conjunctive phrases.

Content & Documentation

find-skills

3891
from openclaw/skills

Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.

General Utilities