Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Chinese-Encoding-Handler
v1.0.0解决PowerShell环境中文文件乱码，支持自动编码检测、安全读写及终端中文显示修复。
⭐ 0· 62·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The SKILL.md and multiple delivery/review documents describe four core scripts (encoding-detector.ps1, safe-read.ps1, safe-write.ps1, terminal-fix.ps1) under a scripts/ directory and show test runs, but the provided file manifest contains only docs and test data—no scripts/ directory or .ps1 script files. That means the package does not actually contain the implementation it claims to provide, which is an incoherence between stated purpose and actual contents.
Instruction Scope
Runtime instructions tell the agent (or user) to execute local PowerShell scripts (e.g., .\scripts\terminal-fix.ps1 -Permanent) which read/write files and can alter terminal/profile settings. Those actions can modify system state and may require administrator privileges. The SKILL.md does not include the code, so following its instructions would require obtaining code from an external source (the SKILL.md references a GitHub repo) — this expands the runtime scope beyond the package and grants the agent discretion to fetch/execute external code.
Install Mechanism
No install spec is provided (instruction-only). Low friction normally, but because the package lacks the scripts it documents, an installer or the agent might be expected to fetch code from the referenced GitHub URL. There is no controlled install URL, release artifact, or integrity information in the package, so any ad-hoc fetch would be higher risk.
Credentials
The skill requests no environment variables or credentials, which is appropriate for a local encoding helper. However, its instructions include making permanent terminal/profile changes (writing to $PROFILE, setting code page) and running operations that may require elevated privileges. Those privileged effects are not declared or constrained in metadata, so the package asks implicitly for system-level changes without proportional disclosure or safeguards.
Persistence & Privilege
Metadata does not request 'always: true' and model invocation is allowed (normal), but the documented terminal-fix operation can persist changes to the user profile or system settings when run with -Permanent (and may require admin rights). A package that makes permanent system-level changes should include its scripts and clear provenance; here the scripts are missing, so the only way to perform persistence would be to fetch/execute external code—this combination increases risk.
What to consider before installing
Do not run any 'terminal fix' or other scripts referenced by this skill until you have the actual script files and have reviewed them yourself. Important points to consider before installing or running anything: 1) The package is documentation and tests only—there are no scripts in scripts/ as the SKILL.md claims. That means an agent or you would need to fetch code from the referenced GitHub repo (or elsewhere) before the skill can work; fetching and executing remote PowerShell without verifying source and integrity is risky. 2) The documented terminal-fix can make permanent changes to your PowerShell profile or system code page and may require administrator privileges—only allow that after inspecting the script and confirming it does only the expected, minimal changes. 3) Ask the publisher for a signed release, a direct install spec (official release URL), or include the .ps1 files in the package; verify file hashes and review the scripts for any network calls, credential exfiltration, or obfuscated logic before running. If the owner supplies the actual scripts and a clear install/release process, re-evaluation could move this to benign; currently the absence of implementation + potential for persistent system modification is why this is suspicious.Like a lobster shell, security has layers — review code before you run it.
latestvk9789fd1wq8jg02gmfx096ss7d83wdmr
License
MIT-0
Free to use, modify, and redistribute. No attribution required.