Walkie-Talkie Mode
AdvisoryAudited by Static analysis on Apr 30, 2026.
Overview
No suspicious patterns detected.
Findings (0)
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.
If the referenced local tools are missing or untrusted, the skill may fail or may run a local program the user did not intend to trust.
The runtime instructions depend on local helper paths/binaries that are not bundled or declared in the supplied install metadata, so users need to ensure those local tools come from trusted sources.
Use `tools/transcribe_voice.sh` ... generate speech using `bin/sherpa-onnx-tts`.
Review or install trusted versions of the transcription and TTS tools, and prefer metadata that declares the required binaries and setup steps.
The skill needs local command execution to create audio replies.
The skill instructs the agent to execute a local TTS command. This is purpose-aligned for local voice generation, but it is still local command execution that should rely on trusted binaries.
bin/sherpa-onnx-tts /tmp/reply.ogg "Tu mensaje aquí"
Use only reviewed local binaries and keep the command paths controlled so the skill does not invoke an unexpected executable.
The skill may send generated text and audio responses to the active WhatsApp conversation.
The workflow includes sending generated files/messages through a messaging tool. This is expected for WhatsApp walkie-talkie mode, but users should understand it can produce outbound messages.
Send the resulting `.ogg` file back to the user as a voice note. ... send `/tmp/reply.ogg` via `message` tool with `filePath`.
Enable it only in intended chats and review sensitive replies or recipients before relying on automatic sending.