中文 EN

Scan Report

Scan New Files

⚠ Low Risk — Risk Score 15/100
Last scan:18 hr ago Rescan
15 /100
voice-clone
多引擎语音合成与声音克隆技能 (Multi-engine TTS and voice cloning)
voice-clone is a legitimate multi-engine TTS tool. One minor shell execution flaw (xdg-open) and an unpinned dependency are present, but no malicious behavior, credential exfiltration, or hidden functionality was found.
Skill Namevoice-clone
Duration38.0s
Enginepi
✓
Safe to install
Review the os.system(xdg-open) call for unnecessary shell execution. Pin dependency versions before production use.

Findings 2 items

Severity Finding Location
Low
Undeclared shell execution for audio playback RCE
The script executes xdg-open via os.system() to auto-play synthesized audio. The output file path is user-influenced, creating a shell injection surface. This is not declared in SKILL.md.
os.system(f"xdg-open '{output_file}' >/dev/null 2>&1 &")
→ Use subprocess.run with shell=False or a platform-appropriate library instead. Remove auto-play entirely if not essential.
 voice-clone.py:224
Low
Dependencies not version-pinned Supply Chain
The install command in SKILL.md installs edge-tts, openai, elevenlabs, coqui-tts, pydantic, aiofiles without version specifiers, risking supply chain attacks from future malicious updates.
pip install edge-tts openai elevenlabs coqui-tts pydantic aiofiles
→ Pin all dependencies to specific versions, e.g. edge-tts==6.1.10, openai>=1.0.0,<2.0.0
 SKILL.md:52
ResourceDeclaredInferredStatusEvidence
Filesystem WRITE WRITE ✓ Aligned voice-clone.py:20 — CACHE_DIR = Path.home() / ".cache" / "voice-clone"
Network NONE READ ✓ Aligned voice-clone.py:65,90,106 — calls to edge-tts, OpenAI API, ElevenLabs API (all le…
Environment NONE READ ✓ Aligned voice-clone.py:72,95 — reads OPENAI_API_KEY and ELEVENLABS_API_KEY, both declare…
Shell NONE WRITE ✓ Aligned voice-clone.py:224 — os.system(f"xdg-open '{output_file}' >/dev/null 2>&1 &")

File Tree

3 files · 11.6 KB · 377 lines
Python 1f · 257L Markdown 1f · 115L JSON 1f · 5L
├─ 📋 _meta.json JSON 5L · 74 B
├─ 📝 SKILL.md Markdown 115L · 2.0 KB
└─ 🐍 voice-clone.py Python 257L · 9.5 KB

Dependencies 6 items

PackageVersionSourceKnown VulnsNotes
edge-tts unpinned pip No Not pinned in SKILL.md install command
openai unpinned pip No Not pinned in SKILL.md install command
elevenlabs unpinned pip No Not pinned in SKILL.md install command
coqui-tts unpinned pip No Not pinned in SKILL.md install command
pydantic unpinned pip No Not pinned in SKILL.md install command
aiofiles unpinned pip No Not pinned in SKILL.md install command

Security Positives

✓ Code is straightforward and readable — no obfuscation, base64, or eval() patterns
✓ Network calls go exclusively to known, legitimate TTS service APIs (Edge, OpenAI, ElevenLabs)
✓ No access to sensitive credential files (.ssh, .aws, .env as files)
✓ No hidden functionality or doc-to-code mismatch beyond the shell execution detail
✓ No credential exfiltration — API keys are used only for intended TTS service calls
✓ No persistence mechanisms (no cron, startup hooks, or backdoor installation)