Skillv1.0.0

ClawScan security

SenseAudio Voice CN · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

ReviewMar 15, 2026, 2:46 AM

Verdict: Review
Confidence: medium
Model: gpt-5-mini
Summary: The skill implements TTS and ASR as described, but its registry metadata omits the required SENSE_API_KEY and the scripts read local OpenClaw config files (including a hardcoded agent path), which is an incoherent / potentially risky mismatch you should verify before installing.
Guidance: What to consider before installing: - The skill will contact https://api.senseaudio.cn for both TTS and ASR and requires a SENSE_API_KEY. The registry entry did not declare this env var — make sure you provide the API key intentionally and understand SenseAudio's account/verification requirements. - The scripts try to read ~/.openclaw/openclaw.json and also a hard-coded path ~/.openclaw/agents/kids-study/openclaw.json to find the API key. Confirm you are comfortable with the skill reading those local config files; if you keep other secrets in those files, review them first. - The code appears to perform only expected operations (HTTP calls to SenseAudio and using edge-tts locally), write audio files to a workspace/audio/ directory, and play audio via local players. There is no obfuscated or clearly malicious code, but the metadata mismatch and hard-coded agent path are red flags. - Recommended actions: inspect the scripts locally, verify or supply only the SENSE_API_KEY (prefer environment variable rather than committing secrets to config files), verify the skill source/author (no homepage provided), and only install if you trust SenseAudio and accept that the skill will access the specified config files and network endpoints. Confidence is medium — the code is coherent with the stated functionality, but the undeclared credential requirement and probing of additional config paths are concrete inconsistencies that warrant caution.
Findings: [no_pre_scan_findings] expected: The repository had no regex-based pre-scan findings. Absence of matches is not evidence of safety, especially for instruction-plus-script skills — manual code review shows the SENSE_API_KEY usage and local file access described above.

Review Dimensions

Purpose & Capability: noteThe code and SKILL.md implement the claimed capabilities (SenseAudio TTS + ASR, and Edge TTS fallback). However the registry metadata lists no required environment variables or credentials even though both SKILL.md and the scripts expect a SENSE_API_KEY. This mismatch between declared requirements and actual needs is inconsistent.
Instruction Scope: concernRuntime instructions and scripts read configuration from ~/.openclaw/openclaw.json and also ~/.openclaw/agents/kids-study/openclaw.json to obtain SENSE_API_KEY. Accessing a hard-coded agent-specific config path (agents/kids-study) is unexpected scope creep — it may expose other user configuration locations and suggests the skill will probe local config files beyond a single, explicitly-declared secret.
Install Mechanism: okThis is instruction-only / script-based with no external binary download. setup.sh installs only the Python 'requests' package and checks for players. package.json lists edge-tts and requests as dependencies, but there's no remote archive or download URL; installation is local via pip. No high-risk download URLs or extract steps were observed.
Credentials: concernThe skill requires a SENSE_API_KEY to use SenseAudio (documented in SKILL.md and enforced in code) but the registry metadata did not declare any required env vars. The scripts also open local openclaw.json files (including another agent's path) which could expose other environment values if present. Requesting a single service API key would be reasonable, but failing to declare it in metadata and probing multiple config paths is disproportionate and surprising.
Persistence & Privilege: okThe skill does not request 'always: true' and does not modify other skills or system-wide agent settings. It writes generated audio files into a workspace audio/ directory (normal for a TTS skill). No elevated or persistent privileges were requested.