← Back to Skills Marketplace
663
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install ardupilot
Description
通过 pymavlink 连接并控制 ArduPilot 无人机。使用此 skill 来操作无人机起飞、降落、移动等。
README (SKILL.md)
ArduPilot 无人机控制 Skill
通过 pymavlink 连接并控制 ArduPilot 无人机 (如 CubeOrange 等)。
⚠️ 重要:起飞核心流程
起飞必须连续发送命令,不要等待!
# 1. 等待飞控稳定 (status=3)
while True:
msg = master.wait_heartbeat(timeout=3)
if msg and msg.system_status == 3:
break
# 2. 连续发送:ARM → GUIDED → TAKEOFF (不要等待!)
master.mav.command_long_send(1, 1, 400, 0, 1, 21196, 0, 0, 0, 0, 0) # ARM (force=21196)
mode_map = master.mode_mapping()
master.set_mode(mode_map['GUIDED']) # GUIDED
master.mav.command_long_send(1, 1, 22, 0, 0, 0, 0, 0, 0, 0, 5) # TAKEOFF 5m
# 3. 监控高度
for i in range(40):
msg = master.recv_match(type='GLOBAL_POSITION_INT', timeout=0.5)
if msg:
alt = msg.relative_alt / 1000
if alt >= 4.5:
print('✅ 到达目标高度!')
break
关键点:
- 发送 ARM 后立即发送 GUIDED 和 TAKEOFF,不要等待
- 如果等待,飞控会重新上锁
- force=21196 是 ArduPilot 的 magic value
连接
from pymavlink import mavutil
master = mavutil.mavlink_connection('tcp:localhost:5762')
master.wait_heartbeat(timeout=10)
system_id = master.target_system # 通常是 1
component_id = master.target_component
1. 检查状态
# 获取飞控状态
msg = master.wait_heartbeat(timeout=5)
print(f'status: {msg.system_status}') # 0=boot, 3=standby, 4=armed
# 获取高度
msg = master.recv_match(type='GLOBAL_POSITION_INT', timeout=1)
print(f'高度: {msg.relative_alt / 1000}m')
# 获取 GPS
msg = master.recv_match(type='GPS_RAW_INT', timeout=1)
print(f'GPS: {msg.satellites_visible}颗, fix={msg.fix_type}')
# 获取电池
msg = master.recv_match(type='SYS_STATUS', timeout=1)
print(f'电池: {msg.voltage_battery / 1000}V')
2. 起飞 (关键流程)
# ⚠️ 必须等待飞控稳定 (status=3)
while True:
msg = master.wait_heartbeat(timeout=3)
if msg and msg.system_status == 3:
break
# ⚠️ 连续发送命令,不要等待!
master.mav.command_long_send(1, 1, 400, 0, 1, 21196, 0, 0, 0, 0, 0) # ARM
mode_map = master.mode_mapping()
master.set_mode(mode_map['GUIDED']) # GUIDED
master.mav.command_long_send(1, 1, 22, 0, 0, 0, 0, 0, 0, 0, 8) # TAKEOFF 8m
# 闭环监控
for i in range(40):
msg = master.recv_match(type='GLOBAL_POSITION_INT', timeout=0.5)
if msg:
alt = msg.relative_alt / 1000
print(f'{i*0.5:.1f}s → {alt:.2f}m')
if alt >= 7.2: # 90% 目标
print('✅ 到达!')
break
3. 降落
# 切换到 LAND 模式
mode_map = master.mode_mapping()
master.set_mode(mode_map['LAND'])
# ⚠️ 必须持续发送 LAND 命令
for i in range(60):
master.mav.command_long_send(1, 1, 21, 0, 0, 0, 0, 0, 0, 0, 0)
import time
time.sleep(0.5)
msg = master.recv_match(type='GLOBAL_POSITION_INT', timeout=0.3)
if msg:
alt = msg.relative_alt / 1000
if alt \x3C 0.3:
print('✅ 降落完成!')
break
4. 相对移动 (LOCAL_POSITION_NED)
# 获取当前位置
local = master.recv_match(type='LOCAL_POSITION_NED', timeout=1)
# X轴前进2米 (NED: X=北)
master.mav.set_position_target_local_ned_send(
0, system_id, component_id,
mavutil.mavlink.MAV_FRAME_LOCAL_NED,
0b0000111111111000,
local.x + 2, local.y, local.z,
0, 0, 0, 0, 0, 0, 0, 0
)
坐标系 (NED)
- X轴: 北 (正=北)
- Y轴: 东 (正=东)
- Z轴: 向下 (负值=向上=高度)
完整示例
from pymavlink import mavutil
import time
master = mavutil.mavlink_connection('tcp:localhost:5762')
master.wait_heartbeat(timeout=10)
print('=== 起飞到 5m ===')
# 1. 等待飞控稳定
while True:
msg = master.wait_heartbeat(timeout=3)
if msg and msg.system_status == 3:
break
print('飞控就绪')
# 2. 连续发送: ARM → GUIDED → TAKEOFF
master.mav.command_long_send(1, 1, 400, 0, 1, 21196, 0, 0, 0, 0, 0)
mode_map = master.mode_mapping()
master.set_mode(mode_map['GUIDED'])
master.mav.command_long_send(1, 1, 22, 0, 0, 0, 0, 0, 0, 0, 5)
print('ARM + GUIDED + TAKEOFF')
# 3. 闭环监控
for i in range(40):
msg = master.recv_match(type='GLOBAL_POSITION_INT', timeout=0.5)
if msg:
alt = msg.relative_alt / 1000
print(f'{i*0.5:.1f}s → {alt:.2f}m')
if alt >= 4.5:
print('✅ 5m!')
break
# === 降落 ===
mode_map = master.mode_mapping()
master.set_mode(mode_map['LAND'])
for i in range(60):
master.mav.command_long_send(1, 1, 21, 0, 0, 0, 0, 0, 0, 0, 0)
time.sleep(0.5)
msg = master.recv_match(type='GLOBAL_POSITION_INT', timeout=0.3)
if msg:
alt = msg.relative_alt / 1000
if alt \x3C 0.3:
print('✅ 降落完成!')
break
注意事项
- 起飞:连续发送命令,不要等待
- 降落:持续发送 LAND 命令
- 闭环检查:每次操作前后获取状态
- GUIDED 模式:自主飞行必须用 GUIDED
- GPS:确保 fix_type >= 3
- 端口:常用 TCP 5762
依赖
pip install pymavlink
Usage Guidance
This skill appears to do what it says: direct control of an ArduPilot vehicle via pymavlink. Before installing or using it: 1) only use it in a safe test environment or simulator first — the instructions explicitly require immediate, repeated control commands and a hard-coded 'force' value that could be unsafe on a live drone; 2) ensure the runtime has access to the local MAVLink endpoint (TCP/serial) and that you trust the source (no homepage/source provided); 3) manually install and verify pymavlink in an isolated environment rather than auto-running unknown code; 4) review and understand each MAVLink command (ARM/TAKEOFF/LAND) and the implications of the '21196' value; and 5) don't run this against a real aircraft until you verify behavior in simulation and have safety cutoffs (kill switch, geofence, spotter).
Capability Analysis
Type: OpenClaw Skill
Name: ardupilot
Version: 1.0.0
The skill bundle is designed for controlling ArduPilot drones via pymavlink. The `SKILL.md` provides clear instructions and Python code snippets for connecting to a local MAVLink instance (`tcp:localhost:5762`), checking drone status, arming, taking off, landing, and performing relative movements. There is no evidence of prompt injection against the AI agent, data exfiltration, malicious execution, persistence mechanisms, or any other harmful behavior. All actions described are consistent with the stated purpose of drone control.
Capability Assessment
Purpose & Capability
Name/description say it controls ArduPilot via pymavlink and the SKILL.md contains only pymavlink-based MAVLink commands and connection details — these are coherent with the stated purpose.
Instruction Scope
Instructions focus on connecting to a local MAVLink endpoint (tcp:localhost:5762), sending ARM/GUIDED/TAKEOFF/LAND and monitoring telemetry; they do not read unrelated files or exfiltrate data. Notes of caution: the guide mandates continuous immediate command sending and uses a hard-coded 'magic' force value (21196), which is operationally significant and potentially unsafe if used on a real vehicle without understanding implications. The SKILL.md also tells the user to run 'pip install pymavlink' even though no install spec is declared in metadata.
Install Mechanism
This is instruction-only (no install spec, no code files) which is low risk from an installation perspective. The README requests pip installing pymavlink, but the skill metadata does not declare this dependency — a minor mismatch to be aware of but not itself malicious.
Credentials
The skill declares no environment variables or credentials, which is appropriate. It does implicitly require network/host access to connect to local MAVLink endpoints (TCP or serial) and permission to install Python packages if the operator follows the instructions.
Persistence & Privilege
always is false and the skill does not request persistent/system-wide privileges or modify other skills. The skill can be invoked autonomously (platform default), which is expected for an actionable control skill.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install ardupilot - After installation, invoke the skill by name or use
/ardupilot - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
- Initial release of the ardupilot skill.
- Provides step-by-step instructions for connecting and controlling ArduPilot drones via pymavlink.
- Covers essential operations: drone status checking, arming, takeoff, relative movement, and landing.
- Emphasizes critical points for reliable takeoff and landing command sequences.
- Includes full usage examples and key safety notes.
Metadata
Frequently Asked Questions
What is ArduPilot Drone Control?
通过 pymavlink 连接并控制 ArduPilot 无人机。使用此 skill 来操作无人机起飞、降落、移动等。 It is an AI Agent Skill for Claude Code / OpenClaw, with 663 downloads so far.
How do I install ArduPilot Drone Control?
Run "/install ardupilot" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is ArduPilot Drone Control free?
Yes, ArduPilot Drone Control is completely free (open-source). You can download, install and use it at no cost.
Which platforms does ArduPilot Drone Control support?
ArduPilot Drone Control is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created ArduPilot Drone Control?
It is built and maintained by LuweiLiao (@luweiliao); the current version is v1.0.0.
More Skills