grandMA2 实体控台控制技能。通过 Telnet 连接控台执行命令。 用于：选灯、调光、存 Cue、播放、查询等 MA2 操作。 触发条件：用户提到 MA2、grandMA2、灯光控台、选灯、存 Cue、执行器等。 执行命令必须使用 ~/ma2_bridge/ma2_cmd.sh 脚本。

Before enabling this skill, confirm the following: (1) You have a trusted ~/ma2_bridge directory and the ma2_cmd.sh / ma2_telnet_server.py scripts are legitimate — the skill relies on those local scripts but does not include them. (2) python3 is required to run the bridge per the docs; ensure python3 is present and that any bridge server you run only connects to your intended MA2 hardware. (3) Set MA2_IP_EXPECTED and MA2_TELNET_PORT_EXPECTED correctly and restrict their values to the known console to avoid accidental commands to other hosts. (4) Because the agent may invoke this skill when certain keywords appear, consider whether you want autonomous invocation enabled for your agent (disable or restrict if you do not want the agent to send commands automatically). (5) If you cannot verify the contents of ma2_cmd.sh and the bridge server, do not enable the skill — a malicious or buggy local script could send unintended commands to lighting hardware.

Type: OpenClaw Skill Name: ma2-control Version: 1.1.0 The skill bundle provides instructions for an AI agent to control grandMA2 lighting consoles via a local bridge service. It is classified as suspicious because it requires high-risk capabilities, including shell execution (bash) and network utilities (nc, curl), and directs the agent to pass user-provided command strings to an external script (~/ma2_bridge/ma2_cmd.sh), which creates a potential shell injection vulnerability. Furthermore, the skill relies on unprovided external scripts and a local service (http://127.0.0.1:40100) that are not included in the bundle, making the complete execution chain unverified.