Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

moviemovie

v1.0.2

搜索电影下载资源(磁力链接/种子),推荐有下载资源的热门电影。 多平台热门聚合(烂番茄+豆瓣)+ 种子资源实时验证 = 只推荐你能下载到的电影。 触发词:"找电影"、"电影下载"、"磁力"、"推荐电影"、"最近有什么好看的"、 "find movie"、"download"、"magnet"、"trending

0· 65·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
medium confidence
Purpose & Capability
Name/description match the code: scripts search torrent indexes, scrape movie charts, build magnet links, and fetch subtitles. Required binary (python3) is appropriate. Minor mismatch: SKILL.md/README assert the skill "will not send user data to third parties," but the skill necessarily issues search queries (movie titles, years, quality prefs) to many external websites/APIs (apibay, bitsearch, torrentdownload, rottentomatoes, douban, subtitlecat, subhd, TorrentClaw/TMDB if keys provided). That claim is therefore incorrect or misleading.
!
Instruction Scope
Runtime instructions and included scripts perform broad web requests and HTML/API scraping across multiple external domains. The SKILL.md instructs the agent to always run subtitle searches in parallel (even if the user didn't ask), and hooks.json/Step 0 run status.py on session start which performs network reachability checks. These behaviors cause automatic outbound queries based on user/session context and are broader than a minimal "on-demand" search; they may leak queried movie names to third parties and run without an explicit user-initiated search.
Install Mechanism
There is no install spec (instruction-only installation), and included Python scripts use only the standard library. No downloads from arbitrary URLs or archive extraction are present. Code is bundled with the skill rather than pulled from an external server at install time.
Credentials
The skill requires no mandatory credentials; two optional env vars (TORRENTCLAW_API_KEY, TMDB_API_KEY) are reasonable for the enhanced features the README describes. However, the README suggests telling the agent/LLM to "save" or configure API keys for you (e.g., paste your key into Claude/OpenClaw to store), which could expose secrets to the model or to platform-managed settings if done carelessly. The code reads only the expected env vars and doesn't appear to access unrelated credentials or config paths.
!
Persistence & Privilege
The skill does not set always:true, but hooks.json runs a status check (python3 scripts/status.py) at SessionStart, causing network reachability checks automatically on session start. While this is consistent with SKILL.md Step 0, auto-running network probes each session increases the skill's silent footprint and may be surprising to users who expect purely on-demand behavior.
What to consider before installing
This skill is functionally coherent for searching torrent/magnet links and subtitles, but consider the following before installing: - Expect outbound network queries: the skill will send your search terms (movie names, years, quality filters) to multiple external sites/APIs (apibay, bitsearch, torrentdownload, rottentomatoes, douban, subtitlecat, subhd, and optionally TorrentClaw/TMDB). The SKILL.md's claim that it "does not send user data to third parties" is inaccurate in practice. - Automatic checks on session start: hooks.json runs a status script every session which performs network checks without explicit user action. If you want strictly on-demand network use, this is a behavior to question. - API keys: TORRENTCLAW_API_KEY and TMDB_API_KEY are optional and appropriate for enhanced features. Do NOT share keys directly into an LLM chat if you want to avoid exposing secrets to model logs; prefer setting environment variables in your agent/platform settings manually. - Legal/privacy: the skill searches and constructs magnet links for torrent resources. Evaluate legal and policy implications for your jurisdiction and hosting environment before use. - Review and sandboxing: because the skill performs scraping and network requests, review the included scripts (they are bundled) and consider running them in a restricted/sandboxed environment if you are concerned about automatic outbound connections. If you accept the above behaviors (network requests, optional API keys, session-start checks), the skill appears to be what it claims; otherwise treat it with caution or ask the author to remove automatic session hooks and to correct misleading statements about data not being sent to third parties.

Like a lobster shell, security has layers — review code before you run it.

latestvk97eyax83v61hjpn6pffszaz7s84aj8m

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Runtime requirements

🎬 Clawdis
Binspython3

Comments