技能 ghe-thread-manager
📦
ghe-thread-manager
内容修订 r1 高风险 ⚙️ 外部命令📁 文件系统访问
管理 GitHub Issue 线程
当 AI 工作在 GitHub issue 和后台任务之间流转时，团队会丢失上下文。此技能可帮助 Claude、Codex 和 Claude Code 跟踪活跃 issue、管理转录状态并创建评审线程。
支持: Claude Codex Code(CC)
安装前请停下并请求确认。
请审查安装计划，并在更改文件前获得用户明确同意。
使用我的智能体安装
将此请求复制给你的智能体，其中包含权威 Skill 页面和清单。
智能体请求
Review the Skillstore skill "ghe-thread-manager" from https://skillstore.io/skills/emasoft-ghe-thread-manager.md and its manifest at https://skillstore.io/api/skills/emasoft-ghe-thread-manager/manifest. Verify the artifact. Stop and obtain explicit user consent before installing or changing files.你的智能体仍应展示其计划，并按安全策略请求必要确认。
Agent 可读资源
当 AI Agent、爬虫或脚本需要干净上下文，而不是读取完整页面时，请使用这些链接。
测试它
正在使用“ghe-thread-manager”。 我们在处理什么？
预期结果:
助手会报告是否已选择某个 issue，并说明转录是开启状态还是私密状态。
正在使用“ghe-thread-manager”。 切换到 issue 42。
预期结果:
助手会更改活跃 issue，宣布转录处于活跃状态，并说明 GitHub 将收到未来的交流内容。
正在使用“ghe-thread-manager”。 实现深色模式。
预期结果:
助手会为该工作创建一个后台 issue，并说明主对话是保持私密，还是停留在当前 issue 上。
安全审计
高风险v8 • 2026/7/6 打开版本化报告
Most static backtick findings are false positives from Markdown code fences and inline command names. Confirmed issues remain for hidden home-directory writes, predictable temporary issue state, public GitHub transcription, absolute prompt-priority language, and shell commands that may receive user-controlled text. The author should tighten permissions, state storage, command invocation, and consent boundaries before publication.
1
已扫描文件
593
分析行数
4
复核项
0
已忽略误报
已确认安全问题 (3)
高
Prompt Injection Attempt Detected
SKILL.md declares an absolute IRON LAW that requires the agent to follow every user word verbatim. This instruction-priority language can conflict with platform policy or security review.
The finding is based on explicit absolute-priority wording in the skill. The language is not framed as optional workflow guidance.
高
Public Conversation Transcription to GitHub
The skill intentionally posts user and assistant exchanges to a GitHub issue when transcription is active. Notifications reduce surprise, but sensitive content can still be exposed.
The skill explicitly states that everything the user says becomes public on GitHub and includes commands to activate issue transcription.
高
User-Controlled Shell Argument Risk
The background thread command template passes title and description placeholders through a shell command. If user text is interpolated without safe argument handling, crafted input could alter command behavior.
The command template clearly includes title and description placeholders. The risk depends on how the agent substitutes user text, so confidence is high enough to retain but not absolute.
能力复核项 (4)
这些是真实的本地能力，对此技能可能是预期行为，因此需要复核，但不会按已确认恶意行为计分。
高
Hidden file in home directory
- ~/.claude (for plugin/settings fixes)
The skill explicitly allows background agents to write to ~/.claude for plugin and settings fixes. That can persistently alter agent configuration outside the project workspace.
高
Hidden file access
- ~/.claude (for plugin/settings fixes)
The skill explicitly allows background agents to write to ~/.claude for plugin and settings fixes. That can persistently alter agent configuration outside the project workspace.
中
Temp directory access
echo "42" > /tmp/ghe_previous_issue
The skill uses a predictable /tmp/ghe_previous_issue path for issue control state. In shared environments, tampering could change which GitHub issue receives transcription.
中
Temp directory access
PREVIOUS=$(cat /tmp/ghe_previous_issue 2>/dev/null)
The skill uses a predictable /tmp/ghe_previous_issue path for issue control state. In shared environments, tampering could change which GitHub issue receives transcription.
风险因素
⚙️ 外部命令 (97)
SKILL.md:53 SKILL.md:54 SKILL.md:76-79 SKILL.md:79-103 SKILL.md:103-105 SKILL.md:105-108 SKILL.md:108-110 SKILL.md:110-123 SKILL.md:123-160 SKILL.md:160-183 SKILL.md:183-194 SKILL.md:194-204 SKILL.md:204-206 SKILL.md:206-215 SKILL.md:215-217 SKILL.md:217-234 SKILL.md:234-240 SKILL.md:240-253 SKILL.md:253-255 SKILL.md:255-266 SKILL.md:266-272 SKILL.md:272-281 SKILL.md:281-283 SKILL.md:283-291 SKILL.md:291-295 SKILL.md:295-312 SKILL.md:312-314 SKILL.md:314-332 SKILL.md:332-338 SKILL.md:338-351 SKILL.md:351-360 SKILL.md:360-364 SKILL.md:364-372 SKILL.md:372-382 SKILL.md:382-384 SKILL.md:384-394 SKILL.md:394 SKILL.md:398-407 SKILL.md:407-411 SKILL.md:411-414 SKILL.md:414-417 SKILL.md:417-420 SKILL.md:420-424 SKILL.md:424-430 SKILL.md:430-433 SKILL.md:433-434 SKILL.md:434-439 SKILL.md:439-441 SKILL.md:441-458 SKILL.md:458-474 SKILL.md:474-475 SKILL.md:475-476 SKILL.md:476-477 SKILL.md:477-478 SKILL.md:478-479 SKILL.md:479-480 SKILL.md:480-481 SKILL.md:481-482 SKILL.md:482-483 SKILL.md:483-484 SKILL.md:484-485 SKILL.md:485-486 SKILL.md:486-494 SKILL.md:494-497 SKILL.md:497-559 SKILL.md:559-560 SKILL.md:560-561 SKILL.md:561-562 SKILL.md:562-563 SKILL.md:563-564 SKILL.md:564-565 SKILL.md:565-566 SKILL.md:566-580 SKILL.md:580-582 SKILL.md:205 SKILL.md:236 SKILL.md:268 SKILL.md:293 SKILL.md:313 SKILL.md:334 SKILL.md:356 SKILL.md:365 SKILL.md:418 SKILL.md:443 SKILL.md:446 SKILL.md:449 SKILL.md:450 SKILL.md:204-206 SKILL.md:234-240 SKILL.md:266-272 SKILL.md:291-295 SKILL.md:312-314 SKILL.md:332-338 SKILL.md:351-360 SKILL.md:364-372 SKILL.md:417-420 SKILL.md:441-458
📁 文件系统访问 (6)
审计者: codex 查看审计历史 →
分享与引用此报告
分享版本化评估报告、中性徽章、嵌入卡片和引用。Skillstore 只报告证据，不判断该 Skill 是否安全。
复制报告链接
https://skillstore.io/skills/emasoft-ghe-thread-manager/audits/8?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdown 徽章
[](https://skillstore.io/skills/emasoft-ghe-thread-manager?utm_source=security_passport_badge)HTML 徽章
<a href="https://skillstore.io/skills/emasoft-ghe-thread-manager?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/emasoft-ghe-thread-manager/security.svg" alt="Skillstore security assessment" loading="lazy"></a>嵌入卡片
<iframe src="https://skillstore.io/embed/skills/emasoft-ghe-thread-manager.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>学术引用 (APA · BibTeX · CFF)
APA 引用
Emasoft. (2026). ghe-thread-manager security audit report (audit version 8) [Author version unspecified]. Skillstore. https://skillstore.io/skills/emasoft-ghe-thread-manager/audits/8BibTeX 引用
@techreport{emasoft-emasoft-ghe-thread-manager-2026,
author = {Emasoft},
title = {ghe-thread-manager security audit report (audit version 8)},
institution = {Skillstore},
year = {2026},
number = {8},
url = {https://skillstore.io/skills/emasoft-ghe-thread-manager/audits/8},
note = {Author version unspecified}
}CITATION.cff
cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "ghe-thread-manager security audit report (audit version 8)"
version: "unspecified"
type: report
authors:
- name: "Emasoft"
date-released: "2026-07-06"
url: "https://skillstore.io/skills/emasoft-ghe-thread-manager/audits/8"
identifiers:
- type: other
value: "skillstore:emasoft-ghe-thread-manager:audit:8"
description: "Skillstore immutable audit report identifier"
Skillstore 评分
评分依据 证据置信度: 中 55
架构
85
可维护性
85
内容
70
社区
87
规范符合性
你能构建什么
切换活跃 Issue 上下文
在 GitHub issue 之间切换，同时保留转录状态和用户通知。
协调后台工作
创建功能或 bug 线程，同时让主对话保持在当前 issue 上。
安全加入评审线程
进入评审 issue，了解新的转录目标，并返回之前的上下文。
试试这些提示
检查当前 Issue
我们现在正在处理哪个 issue？
切换到某个 Issue
切换到 issue 42，并告诉我转录是否处于活跃状态。
创建后台工作
为添加深色模式创建一个后台功能线程，但让此聊天保持当前转录状态。
恢复并继续上下文
找到上一个活跃 issue，总结它，并在恢复转录前询问我。
最佳实践
- 在讨论私密数据或客户数据之前，确认活跃 issue。
- 在分享密钥或敏感上下文之前，使用 go private 或 stop tracking。
- 在启用此技能之前，确保 GHE 辅助脚本已完成评审并固定版本。
避免
- 加入或切换 issue 后，不要假设转录是私密的。
- 不要在没有目标仓库 GitHub 权限的情况下使用此技能。
- 不要让后台 issue 创建改变主对话状态。
常见问题
此技能会向 GitHub 发布消息吗？
会，当转录处于活跃状态时。它需要提供通知，说明活跃 GitHub issue 目标。
我可以停止转录吗？
可以。要求 stop tracking、go private 或 stop transcribing 即可关闭转录。
创建后台线程会切换我的活跃 issue 吗？
不会。后台功能和 bug 线程独立于主对话状态。
此技能需要哪些工具？
它需要 GHE 辅助脚本、有效的插件根目录，以及用于查找 issue 的 GitHub CLI 访问权限。
它能找到上一个活跃 issue 吗？
可以。它可以检查已保存的状态，并且可以搜索最近的 GitHub issue 作为备用方案。
它适合处理密钥吗？
不安全。请将活跃转录视为公开发布到 GitHub，并在分享密钥之前停止跟踪。