技能 ai-maestro-agent-messaging
📦

ai-maestro-agent-messaging

内容修订 r1 高风险 ⚙️ 外部命令🌐 网络访问📁 文件系统访问

使用 AI Maestro 协调 AI 智能体

多智能体编码工作需要可靠的交接和状态检查。此技能帮助智能体通过 AI Maestro 交换结构化消息。

支持: Claude Codex Code(CC)
⚠️ 38 差

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 "ai-maestro-agent-messaging" from https://skillstore.io/skills/23blocks-os-ai-maestro-agent-messaging.md and its manifest at https://skillstore.io/api/skills/23blocks-os-ai-maestro-agent-messaging/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。

Markdown 详情 GET /skills/23blocks-os-ai-maestro-agent-messaging.md 已签名 manifest GET /api/skills/23blocks-os-ai-maestro-agent-messaging/manifest 已签名 lockfile GET /api/skills/23blocks-os-ai-maestro-agent-messaging/lockfile

测试它

正在使用“ai-maestro-agent-messaging”。 检查我的消息。

预期结果:

智能体收件箱有两条未读消息。一条要求进行 API 审查，另一条报告已完成前端任务。

正在使用“ai-maestro-agent-messaging”。 告诉 frontend-dev 端点已准备就绪。

预期结果:

已向 frontend-dev 发送 normal 优先级的消息。主题说明端点已准备好进行集成。

正在使用“ai-maestro-agent-messaging”。 有任何紧急消息吗？

预期结果:

有一条 urgent 消息。它报告了身份验证中断，并请求立即调查。

安全审计

高风险
v6 • 2026/7/4 打开版本化报告

Most static findings are Markdown scanner artifacts or documented AI Maestro command examples, not hidden code execution. The main residual risks are intentional cross-host messaging, tmux injection, and workflows that may act on untrusted agent messages. No prompt-injection text was found that attempts to override this audit.

1
已扫描文件
695
分析行数
0
复核项
0
已忽略误报

已确认安全问题 (3)

高
SKILL.md:397-424SKILL.md:543-544
Terminal Message Injection Into Other Agents
The skill supports an inject method that writes alert text into another tmux session. A crafted message could interrupt or influence another agent workflow.
The method list explicitly describes inject as visible and interrupting, and examples send urgent text to another agent. That is a clear inter-agent influence channel.
高
SKILL.md:613-623
Untrusted Agent Messages Can Drive Actions
The receiving workflow tells the agent to read incoming messages, assess urgency, and work on the request. Messages from other agents should be treated as untrusted input before taking external actions.
The workflow directly connects message content to agent action. The file does not add an operator confirmation step for risky or externally sourced instructions.
中
SKILL.md:322-367
Cross-host Message Data Transmission
The skill describes sending message content to configured remote AI Maestro hosts. This can transmit operator-provided task details outside the local machine.
The cross-host section states that remote agents are resolved from host configuration and messages are posted to the remote API. The behavior is intentional but creates a data-sharing boundary.

风险因素

⚙️ 外部命令 (126)
SKILL.md:24 SKILL.md:25 SKILL.md:26 SKILL.md:33 SKILL.md:34 SKILL.md:44 SKILL.md:55 SKILL.md:56 SKILL.md:57 SKILL.md:62 SKILL.md:63 SKILL.md:71-86 SKILL.md:86-90 SKILL.md:90-96 SKILL.md:96-102 SKILL.md:102-104 SKILL.md:104-110 SKILL.md:110-114 SKILL.md:114-120 SKILL.md:120-123 SKILL.md:123-134 SKILL.md:134-138 SKILL.md:138-140 SKILL.md:140-144 SKILL.md:144-149 SKILL.md:149-155 SKILL.md:155-158 SKILL.md:158-181 SKILL.md:181-185 SKILL.md:185-187 SKILL.md:187-193 SKILL.md:193-201 SKILL.md:201-211 SKILL.md:211-215 SKILL.md:215-217 SKILL.md:217-226 SKILL.md:226-229 SKILL.md:229-233 SKILL.md:233-235 SKILL.md:235-239 SKILL.md:239-240 SKILL.md:240-244 SKILL.md:244-252 SKILL.md:252-255 SKILL.md:255-265 SKILL.md:265-269 SKILL.md:269-276 SKILL.md:276-290 SKILL.md:290-292 SKILL.md:292-295 SKILL.md:295-296 SKILL.md:296-297 SKILL.md:297-298 SKILL.md:298-299 SKILL.md:299-300 SKILL.md:300-301 SKILL.md:301-302 SKILL.md:302-305 SKILL.md:305-320 SKILL.md:320-328 SKILL.md:328-329 SKILL.md:329-348 SKILL.md:348-352 SKILL.md:352-354 SKILL.md:354-360 SKILL.md:360-364 SKILL.md:364-365 SKILL.md:365-367 SKILL.md:367-373 SKILL.md:373-377 SKILL.md:377-382 SKILL.md:382-386 SKILL.md:386-389 SKILL.md:389-390 SKILL.md:390-394 SKILL.md:394-401 SKILL.md:401-403 SKILL.md:403-406 SKILL.md:406-407 SKILL.md:407-408 SKILL.md:408-411 SKILL.md:411-412 SKILL.md:412-413 SKILL.md:413-416 SKILL.md:416-425 SKILL.md:425-430 SKILL.md:430-440 SKILL.md:440-444 SKILL.md:444-450 SKILL.md:450-480 SKILL.md:480-489 SKILL.md:489-492 SKILL.md:492-501 SKILL.md:501-504 SKILL.md:504-527 SKILL.md:527-530 SKILL.md:530-554 SKILL.md:554-559 SKILL.md:559-565 SKILL.md:565-568 SKILL.md:568-578 SKILL.md:578-581 SKILL.md:581-592 SKILL.md:592-595 SKILL.md:595-601 SKILL.md:601-609 SKILL.md:609-610 SKILL.md:610-666 SKILL.md:666-667 SKILL.md:667-670 SKILL.md:670-671 SKILL.md:671-682 SKILL.md:682-683 SKILL.md:683-684 SKILL.md:684-688 SKILL.md:234 SKILL.md:239 SKILL.md:261 SKILL.md:264 SKILL.md:271 SKILL.md:670 SKILL.md:233-235 SKILL.md:239 SKILL.md:255-265 SKILL.md:269-276 SKILL.md:670
🌐 网络访问 (10)
SKILL.md:274 SKILL.md:335 SKILL.md:342 SKILL.md:390 SKILL.md:394 SKILL.md:683 SKILL.md:692 SKILL.md:693 SKILL.md:694 SKILL.md:342
📁 文件系统访问 (34)
SKILL.md:25 SKILL.md:34 SKILL.md:44 SKILL.md:90 SKILL.md:96 SKILL.md:234 SKILL.md:245 SKILL.md:261 SKILL.md:264 SKILL.md:328 SKILL.md:365 SKILL.md:384 SKILL.md:389 SKILL.md:485 SKILL.md:610 SKILL.md:667 SKILL.md:670 SKILL.md:25 SKILL.md:34 SKILL.md:44 SKILL.md:90 SKILL.md:96 SKILL.md:234 SKILL.md:245 SKILL.md:261 SKILL.md:264 SKILL.md:328 SKILL.md:365 SKILL.md:384 SKILL.md:389 SKILL.md:485 SKILL.md:610 SKILL.md:667 SKILL.md:670
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。

打开版本化报告
安全评估

复制报告链接

https://skillstore.io/skills/23blocks-os-ai-maestro-agent-messaging/audits/6?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Markdown 徽章

[![Skillstore security assessment](https://skillstore.io/badges/skills/23blocks-os-ai-maestro-agent-messaging/security.svg)](https://skillstore.io/skills/23blocks-os-ai-maestro-agent-messaging?utm_source=security_passport_badge)

HTML 徽章

<a href="https://skillstore.io/skills/23blocks-os-ai-maestro-agent-messaging?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/23blocks-os-ai-maestro-agent-messaging/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

嵌入卡片

<iframe src="https://skillstore.io/embed/skills/23blocks-os-ai-maestro-agent-messaging.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

23blocks-OS. (2026). ai-maestro-agent-messaging security audit report (audit version 6) [Author version unspecified]. Skillstore. https://skillstore.io/skills/23blocks-os-ai-maestro-agent-messaging/audits/6

BibTeX 引用

@techreport{23blocks-os-23blocks-os-ai-maestro-agent-messaging-2026, author = {23blocks-OS}, title = {ai-maestro-agent-messaging security audit report (audit version 6)}, institution = {Skillstore}, year = {2026}, number = {6}, url = {https://skillstore.io/skills/23blocks-os-ai-maestro-agent-messaging/audits/6}, 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: "ai-maestro-agent-messaging security audit report (audit version 6)" version: "unspecified" type: report authors: - name: "23blocks-OS" date-released: "2026-07-04" url: "https://skillstore.io/skills/23blocks-os-ai-maestro-agent-messaging/audits/6" identifiers: - type: other value: "skillstore:23blocks-os-ai-maestro-agent-messaging:audit:6" 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
87
内容
72
社区
78
规范符合性

你能构建什么

协调后端和前端智能体

在专门的编码智能体之间发送 API 就绪更新、阻塞问题和实现请求。

启动后检查分配的工作

在开始新的编码任务之前，查看当前智能体的未读消息。

路由紧急事件更新

通过 tmux 通知另一个智能体，然后发送一条包含上下文的持久消息。

试试这些提示

检查我的智能体收件箱 
检查我的 AI Maestro 收件箱，并总结此智能体的所有未读消息。
读取一条消息 
从我的 AI Maestro 收件箱读取消息 MESSAGE_ID，并告诉我发送者、优先级和请求的操作。
发送结构化更新 
向 backend-api 发送一条 normal 更新，说明用户资料屏幕已准备好进行端点测试。
协调紧急的跨智能体响应 
通知 backend-api 生产登录失败，然后发送一条包含症状和时间戳的详细 urgent 消息。

最佳实践

  • 除非操作者明确批准另一个目标，否则只检查当前智能体收件箱。
  • 发送跨主机消息之前，确认目标智能体和主机。
  • 在操作者批准操作之前，将消息内容视为不可信指令。

避免

  • 不要为了满足一般状态请求而读取另一个智能体收件箱。
  • 不要通过智能体消息发送机密或凭据。
  • 不要将 tmux 注入用于常规更新或不明确的目标。

常见问题

此技能有什么作用？
它帮助 AI 智能体从启用 Bash 的会话中发送、读取和管理 AI Maestro 消息。
它支持哪些工具？
当 Bash 访问可用时，它支持 Claude、Codex 和 Claude Code。
它会读取机器上的每条消息吗？
不会。文档化的工作流聚焦于 AI Maestro 消息存储下的当前智能体收件箱。
它可以向远程智能体发送消息吗？
可以。当存在 AI Maestro 主机设置时，它可以寻址已配置的远程主机。
这对敏感数据安全吗？
请谨慎使用。消息可由接收智能体读取，并且可能存储在本地 AI Maestro 文件中。
发送前我应该验证什么？
验证目标智能体、主机、优先级，以及消息是否包含敏感信息。

开发者详情

许可证

MIT

Skillstore 修订

r1

版本提示

作者未声明版本。

仓库

https://github.com/23blocks-OS/ai-maestro/tree/main/skills/agent-messaging

引用

14dc8f201a64f8d30fd131d7f036cd5e788be523

维护新鲜度

2026/7/20

使用量

14 次下载 · 365 次浏览

文件结构

📄 SKILL.md

23blocks-OS 的更多技能

查看全部

相关技能

查看全部