Speak Turbo - Talk to your Claude 90ms latency!
PassAudited by ClawScan on May 1, 2026.
Overview
This appears to be a straightforward local text-to-speech skill, with the main things to notice being its package install/build steps, local background daemon, and WAV file output.
Before installing, be comfortable with a local TTS daemon that starts in the background, package installation/build steps, and audio file writes in allowed directories. The artifacts do not show credential use, data exfiltration, or deceptive behavior.
Findings (3)
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.
Installing it may change the user's Python environment and add a new executable under `~/.local/bin`.
The install script pulls external packages with version ranges, builds a local binary, and places it in the user's PATH. This is normal for a TTS CLI but is a supply-chain and local-environment change users should notice.
pip install --quiet "pocket-tts>=0.1.0,<1.0" ... cargo build --release --quiet cp target/release/speakturbo ~/.local/bin/
Install only from a trusted source, prefer a virtual environment if possible, and pin or review dependencies if strict supply-chain control is needed.
The skill may keep a local TTS server running in the background and consuming memory for up to an hour after use.
The CLI starts a detached local daemon and records its PID, so a speech request can leave a background process running. SKILL.md discloses the daemon behavior and one-hour idle shutdown.
process = subprocess.Popen(
[sys.executable, str(daemon_script)],
stdout=log,
stderr=log,
start_new_session=True,
)
PID_FILE.write_text(str(process.pid))Use it only if a local background TTS daemon is acceptable, and use the documented daemon stop command or `pkill -f "daemon_streaming"` when you want it stopped.
An agent using the skill can save audio files in allowed directories, and can write elsewhere if the user explicitly allows another directory.
The tool can create WAV files and lets users expand the writable directory set. The allowlist and override mechanisms are disclosed and scoped, so this is a purpose-aligned note rather than a concern.
The `-o` flag only writes to directories that are on the allowlist... If you need to write elsewhere, use `--allow-dir`... To permanently allow a directory, add it to `~/.speakturbo/config`
Use explicit output paths and avoid permanently allowing broad directories such as the entire home directory unless you intend that access.