Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Python包鸿蒙兼容性测试技能
v1.0.1Check Python library compatibility with HarmonyOS. Downloads source from GitHub/PyPI, detects Windows-specific dependencies, runs pytest with per-test-case r...
⭐ 0· 344·0 current·0 all-time
byterry.King@terryking1992
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
Name/description align with the code and SKILL.md: the script downloads package sources, scans for Windows imports, discovers and runs pytest tests, and writes reports. Small metadata inconsistency: registry metadata lists version 1.0.1 while manifest.json contains version 1.2.0 (minor mismatch that should be resolved). Overall functionality is coherent with the stated purpose.
Instruction Scope
Runtime instructions and the included script instruct the agent to download archives from GitHub/PyPI, extract them, install packages and run pytest test suites. Running tests implies executing untrusted third-party code (package test suites) which can perform arbitrary actions on the host — this is expected for a tester but is high-risk if not sandboxed. Additionally, the script explicitly disables SSL certificate verification when downloading from GitHub (ssl.CERT_NONE), which weakens transport security and could enable man-in-the-middle manipulation of the downloaded source.
Install Mechanism
No install spec (instruction-only skill) — low friction. However, the runtime code performs downloads and archive extraction to temporary directories and may install packages via pip/pytest at runtime. Those runtime actions write files and execute code; ensure they run in a confined environment.
Credentials
The skill does not request environment variables, credentials, or config paths. The lack of declared sensitive requirements is appropriate for its purpose. Note: because it downloads and executes third-party code, it should not be run in environments that expose secrets or privileged mounts.
Persistence & Privilege
always is false and the skill is user-invocable; it doesn't request persistent system-level privileges in the manifest. It writes reports and may keep downloaded sources if --keep-source is used; that is expected and limited to its own workspace.
What to consider before installing
This skill does what it says (downloads package source, scans for Windows imports, runs tests and emits reports) but has two practical security concerns you should consider before installing or running it:
1) TLS bypass: The included script disables TLS certificate verification when downloading from GitHub/PyPI. That is unnecessary in normal environments and makes downloads susceptible to man-in-the-middle tampering. Ask the author to remove disabling of ssl verification (restore default SSL checks) or explain why it is required for your environment.
2) Execution of untrusted code: To assess compatibility the tool downloads packages and runs their test suites. Tests can execute arbitrary Python code (including network access, file operations, or deleting files). Only run this tool in an isolated sandbox (container, VM, or dedicated build runner) that has no access to secrets, credentials, or sensitive mounts. Prefer ephemeral environments; do not run on developer workstations or production hosts.
Other points to verify:
- Resolve the version mismatch between registry metadata (1.0.1) and manifest.json (1.2.0).
- If you need safer operation, request changes: re-enable SSL verification, add optional allowlist of trusted repos/packages, add explicit safeguards (no-network mode, resource/time limits, capability drops), and document exactly which subprocesses (pip, pytest) are invoked.
- Use --keep-source only for manual inspection in a safe place; retained sources may contain malicious code.
If the author provides evidence that TLS is only disabled for a narrow, documented HarmonyOS reason or they add sandboxing/allowlisting, my assessment would move toward benign. Until then, treat this skill as potentially risky but not clearly malicious.Like a lobster shell, security has layers — review code before you run it.
latestvk975tww214fzm2zt7948ts83n582kyk9
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
