Pi Speaker
PassAudited by ClawScan on May 1, 2026.
Overview
Pi Speaker appears to do what it claims: generate or use an audio file and play it on the gateway host speaker, with the main caveat that it relies on narrowly scoped local exec/audio commands.
This skill is reasonable to install if you want the agent to speak through the gateway host speaker. Before using it, make sure you are comfortable allowing the agent to run the specific local playback command, avoid unnecessary elevated shell access, and remember that spoken announcements may be audible to people nearby.
Findings (2)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
The agent can cause audio to play on the gateway host speaker and will run a local playback command to do so.
The skill directs the agent to run a local host command. That is disclosed and necessary for local audio playback, but it remains a sensitive capability that should stay scoped to the listed playback commands.
Call the **exec** tool with exactly one of: `pw-play <path>`, `paplay <path>`, or `$HOME/bin/openclaw-speaker-play.sh <path>`.
Use only the listed paplay, pw-play, or helper-script commands; avoid elevated shells unless truly required; and quote or otherwise safely pass file paths, especially for any user-supplied audio file path.
A successful exec result may mean playback was launched, not necessarily that the whole audio file finished playing.
The helper backgrounds the audio player and exits. This is understandable for non-blocking playback, but it means the process may continue briefly after the tool call returns.
nohup paplay "$path" >/dev/null 2>&1 & ... nohup pw-play "$path" >/dev/null 2>&1 &
Keep audio files short for announcements and be aware that stopping playback may require terminating the paplay or pw-play process if something unexpected is played.