Security audit
Security checks for vulnerabilities and agentic risk
This skill is a straightforward Bilibili favorites downloader; it uses network access, yt-dlp, and local file writes in ways that match its stated purpose, with storage and bandwidth caveats users should understand.
Install only if you want an agent to run a local Python downloader that contacts Bilibili, invokes yt-dlp, and saves videos to disk. Use --list-only first, then set an explicit output directory, size limit, and max-download value to avoid unexpected bandwidth or storage use.
The skill explicitly drives network access and shell execution via a Python downloader and yt-dlp, but it does not declare any tool scope or permission boundaries. In an agent setting, this weakens security controls and user transparency, increasing the chance the skill can perform broader-than-expected actions such as network fetching and local command execution without clear policy gating.
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
pn: 页码
ps: 每页数量
"""
url = f"https://api.bilibili.com/x/v3/fav/resource/list?media_id={fid}&pn={pn}&ps={ps}&keyword=&order=mtime&type=0&tid=0&platform=web"
headers = {
'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36',
subprocess module calls execute external commands. Without careful input validation, this enables command injection.
]
try:
result = subprocess.run(cmd, capture_output=True, text=True, timeout=30)
if result.returncode == 0:
info = json.loads(result.stdout)
return info
subprocess module calls execute external commands. Without careful input validation, this enables command injection.
print(f"BV号: {bvid}")
try:
result = subprocess.run(cmd, capture_output=False, text=True)
if result.returncode == 0:
print(f"✅ 下载成功: {title}")
return True, "downloaded"
The README makes clear that the tool downloads videos, but it does not prominently warn that running the skill can trigger bulk network transfers and persistent writes to the user's local filesystem. In an agent/skill context, this matters because users may invoke the skill conversationally without realizing it will save many files, consume disk space and bandwidth, and potentially overwrite organizational expectations about passive assistance versus active file operations.
The skill instructs downloading potentially many videos to local storage and even promotes bulk usage, but it does not clearly warn users about disk consumption or that files will be written to the local filesystem. In practice this can lead to unexpected storage usage, clutter, or accidental writes in sensitive environments, especially because the workflow defaults to downloading rather than listing only.
The module docstring presents the skill's purpose and behavior only in Chinese. Under the policy rule for language/locale, forcing a single language without user opt-in can be a natural-language policy violation when no alternative or justification is provided.
No suspicious patterns detected.