doubao-tts
Analysis
This looks like a straightforward text-to-speech skill, but it requires your Volcengine token and sends the text you ask it to read to ByteDance/Volcengine to generate audio.
Findings (4)
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.
Checks for instructions or behavior that redirect the agent, misuse tools, execute unexpected code, cascade across systems, exploit user trust, or continue outside the intended task.
output_file="${save_path:-/tmp/doubao_tts_$(date +%s).mp3}" ... base64 -d > "$output_file" ... afplay "$output_file"The skill writes decoded audio to a local file and invokes a local audio player. This is appropriate for TTS playback, but it is still local file and tool use.
"repository": { "type": "git", "url": "" } ... "dependencies": ["jq", "curl", "base64"]The package metadata has no repository URL and documents external helper tools. This limits provenance information but does not show hidden or unsafe installation behavior.
Checks whether tool use, credentials, dependencies, identity, account access, or inter-agent boundaries are broader than the stated purpose.
access_token=$(cat "$config_file" | jq -r '.access_token') ... -H "Authorization: Bearer;$access_token"
The skill reads a local Volcengine access token and uses it to authorize TTS API calls. This is expected for the stated service integration, but the token may grant account usage or incur provider charges.
Checks for exposed credentials, poisoned memory or context, unclear communication boundaries, or sensitive data that could leave the user's control.
api_url="https://openspeech.bytedance.com/api/v1/tts" ... "uid": "388808087185088" ... "text": "$text"
The skill sends the requested text, along with a fixed user identifier field, to the external ByteDance/Volcengine TTS endpoint. This is core to the skill's purpose, but it means text content leaves the local machine.
