senseaudio-floating-audio-assistant

Use when the user wants to open, stop, configure, debug, or package the SenseAudio floating audio assistant in AudioClaw, including system-audio subtitles, b...

This package appears coherent and implements what it claims: macOS overlay + SenseAudio integration. Before installing, do the following: (1) Ensure you trust the SenseAudio service (api.senseaudio.cn) because live TTS will send your API key there; the smoke script only contacts SenseAudio when --live-tts is used. (2) Inspect the external workspace tools referenced by TOOL_DIR (workspace/tools/realtime_interpreter) — the start/stop/runner scripts that actually perform capture/overlay are not included in this upload and could perform arbitrary actions. (3) Provide the SENSEAUDIO_API_KEY (or AUDIOCLAW_ASR_API_KEY) via your runtime environment or workspace .env; the scripts will check for but not leak the key. (4) Be aware the skill will run swiftc (compile a native overlay) and execute shell scripts that alter macOS audio routing; run in a safe/testing environment first if you are unsure. Overall, the skill is internally consistent but verify the external toolchain before execution.

Type: OpenClaw Skill Name: senseaudio-floating-audio-assistant Version: 1.0.0 The skill bundle is a legitimate management interface for a macOS audio assistant integration. It provides wrapper scripts (e.g., start-senseaudio-floating-audio-assistant.sh) to control an external audio processing toolchain, manage macOS audio routing via SwitchAudioSource, and interact with the SenseAudio API (api.senseaudio.cn) for ASR and TTS. The bundle includes robust diagnostic tools like doctor-senseaudio-floating-audio-assistant.sh and senseaudio_api_smoke.py, which specifically check for the presence of required dependencies and ensure that sensitive credentials are not accidentally exposed or printed. No evidence of malicious intent, data exfiltration, or unauthorized persistence was found.