Skills/ClawHub/win-dev-pitfalls

win-dev-pitfalls@1.71.0

CAUTION· install only

Publisher claims: "Windows development pitfalls vault — bat/cmd batch, PowerShell 5.1/7, Win32 GUI, Flutter desktop, cross-platform pwsh. Encoding traps (GBK/BOM/CRLF), cmd syntax gotchas, UAC elevation, sandbox testing, delivery checklists. Windows 开发踩坑经验库：批处理、PowerShell、Win32 GUI、Flutter 桌面端适配与审计。Keywords: bat, cmd, PowerShell, ps1, pwsh, BOM, CRLF, 闪退, 乱码, Flutter, Dart, Win32, Windows desktop, Windows scripting" Observed in sandbox: no network, touches nothing outside the project, writes no memory.

Registry ClawHubPublisher mowenqwq · publishing for 81 daysInstalls 564Last detonated 8d ago
What would fix thisWatchDispute verdict

Coverage

Nothing executable in the bundle ran; the trace covers the install step only.
Tested
every file read, file written and host contacted by what ran
Not tested
use of the skill by an agent

Is [email protected] safe to install?

KillerSkills ran it in the KillerSkills honeypot sandbox (1 scenario; nothing executable in the bundle ran; the trace covers the install step only) on 2026-11-07 and scored it 77/100 (Caution). It is not known to be malicious, but it has behaviour or code a security policy should decide on before install. Notable findings: Destructive file system command detected; Hidden zero-width characters detected; Hidden zero-width characters detected. Summary: no network, touches nothing outside the project, writes no memory.

Behavior trace

Honeypot run · 0 s · 1 scenarios · 0 deviations
t+CategoryObserved action
0.0sprocessno install scripts in bundle
0.0sprocessscript backend runs setup only
0.0sprocessscript backend runs setup only
0.0sprocessno install scripts to re-run

Static findings

MEDIUM
Process kill commandSKILL.md:756 · Process kill command — agent can terminate processes- **v1.16.0 (2026-08-18)**:§5/§7 代码清理实战新坑(对应一次"移除内核驱动加载路径、仅保留关闭记录逻辑"的大改): ① **删变量/函数后必须 grep 残留引用(静默失效)**: 移除 `$script:ExtremeActive` 变量定义后, kill 分支 `if (-not $script:ExtremeActive)` 仍在引用它——PS 未定义变量返回Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Python subprocess execution detectedSKILL.md:746 · Python subprocess execution detected- **v1.44.0 (2026-08-22)**:**新增「EDR-Freeze 反冻防御(用户态纵深) + 全源码漏洞审计」实战(对应 SilverFox Detector v2.15.8)**:① **攻击面**: 攻击者滥用 WER(`WerFaultSecure.exe` 调 `MiniDumpWriteDump`)在转储竞态窗口用 `SuspendThread`/`NtSuspendFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
HIGH
Destructive file system command detectedSKILL.md:815 · Destructive file system command detectedif exist "%TEMP%\ver.txt" ( set /p VER=<"%TEMP%\ver.txt" & del /q "%TEMP%\ver.txt" )Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Process kill commandSKILL.md:751 · Process kill command — agent can terminate processes- **v1.35.0 (2026-08-21)**:**架构变更:自保护从 PowerShell 层下沉到 Go exe**(对应 SilverFox Detector **v2.15.0**):① **变更实质**: 原 §5/§7 记录的 `SelfGuard.ps1`(看门狗+进程 DACL) 与 `SilverFoxDetect.ps1::Test-Integrity`(文件完整性) 全Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
HIGH
Hidden zero-width characters detectedSKILL.md:382 · Hidden zero-width characters detected — may conceal content**根因**: 回写脚本 `"" + text` 前置 BOM, 而解码时用了 `decode('utf-8')`(未剥原 BOM → text 已含 `\ufeff` 字符) → 文件头变成**双重 BOM**(`EF BB BF EF BB BF`)。PowerShell 5.1 读取时剥离一个 BOM, 剩余 `\ufeff` 位于首行注释前, 导致脚本被**静默拒绝执行**(无报错、无输Fix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.
AST07
MEDIUM
File type combined with upload actionSKILL.md:764 · File type combined with upload action — agent may exfiltrate local data files- **v1.9.0 (2026-08-17)**:§8/§5 第五轮深度检测新坑: ① **清单/Flag 行的"路径+[状态]"双段格式是反复踩的坑**: 所有从 `[高危文件][原因] 路径 [状态]` 行提取路径的地方, 必须**先 -replace '\s+\[[^\[\]]*\]\s*$' 剥尾部状态**再取(贪婪 (.+)$ 会把 [建议隔离] 吞进路径, 实测 /interactiFix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.
AST07
HIGH
Hidden zero-width characters detectedSKILL.md:1 · Hidden zero-width characters detected — may conceal content---Fix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.
AST07
MEDIUM
File permission modification commandSKILL.md:772 · File permission modification command — agent changes file ownership/permissions- **v1.4.3 (2026-08-17)**:§5 新增威胁锁定实战经验: ① **PowerShell 管道 0 对象不会触发 Set-Content** → 清空文件用 `Clear-Content`, 否则 `$lines | Set-Content` 空数组时文件保持原样(清单删除静默失败!); ② 文件锁定: Linux `chattr +i`(immutable 连 root 也Fix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.
AST05
MEDIUM
Process kill commandSKILL.md:770 · Process kill command — agent can terminate processes- **v1.5.0 (2026-08-17)**:§7 新增自查方法论 + §2/§5 补坑: ① **多 agent 并行代码审查** + 逐个复验(不轻信结论, 用 grep/sed 验证每个发现) 是发现隐蔽 bug 的高效方式; ② **`foreach ($a in $args)` 循环内 `$_` 是空**(foreach 不设 `$_`, 只有 ForEach-Object 管道才Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Process kill commandSKILL.md:762 · Process kill command — agent can terminate processes- **v1.11.0 (2026-08-17)**:§1/§2/§3/§5/§8 第七轮用户实测反馈修复(v1.43~v1.47)新坑: ① **`Console.OutputEncoding = GetEncoding(936)` 与 bat `chcp 65001` 冲突=退出乱码**: PS 子进程输出按 GBK 编码字节, cmd 按 UTF-8 解码 → 退出段满屏符号乱码(♥♡❤☺♦Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Process kill commandSKILL.md:869 · Process kill command — agent can terminate processes- **进程防杀三层(v1.44 实战)**:① 进程 DACL 移除 Everyone/Users 的 PROCESS_TERMINATE(非管理员 taskkill → Access Denied,`SetKernelObjectSecurity` + SDDL `O:SYG:SYD:(A;;GA;;;SY)(A;;GA;;;BA)`);② **watchdog 双进程守护**: 引擎被杀自动Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05

Cost to the agent

Instructions loaded
≈31,036 tok
1 file · 2 in bundle

Tokens estimated at four characters each. Skills load their instruction text when triggered; the figure is what that costs.

Known vulnerabilities

Vulnerability advisories (OSV) cover npm and PyPI packages; this registry has no advisory database.

What would make this Verified

3 points to 80 · weights: behavior 45%, static 25%, prompt 15%, publisher 15%
  1. Unclaimed publisher≈ +8 on the totalClaim mowenqwq: a verified publisher gets +50 on publisher trust. It is free and usually instant with GitHub sign-in.
  2. Destructive file system command detected≈ +4 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.
  3. Hidden zero-width characters detected≈ +4 on the totalKeep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.
  4. Process kill command≈ +2 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.
  5. Python subprocess execution detected≈ +2 on the totalBuild commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
  6. File type combined with upload action≈ +2 on the totalRemove the call, or declare the host in frontmatter and explain in the README what is sent and why.
Publish a new version to ClawHub, then queue it here. Not your skill? Claim the handle to get verdict changes by email.Claim mowenqwqPropose a patch to the maintainer