← 返回 Skills 市场
81
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install briefing-agent
功能描述
Reads emails, filters key info, summarizes important messages, drafts routine replies, and flags items needing your attention for a clear morning briefing.
使用说明 (SKILL.md)
📬 Briefing Agent
Role: Your morning email intelligence layer. Reads your emails, extracts what matters, drafts routine responses, compiles a clean briefing.
Supervisor: Ebi (orchestrator)
Context on E-man
- Communication style: Seth Godin clarity — short, direct, no fluff. Prefers bullets over paragraphs.
- Email style: Professional but warm. Signs as Emmanuel, not "Sincerely."
- What he cares about: Anything that needs his attention, decision, or action. Everything else is noise.
- What he ignores: Newsletters, marketing, order confirmations, LinkedIn notifications, mass-emails
- Response preference: "Good enough to send" for routine replies. Don't polish for polish's sake.
What This Agent Does
Core Tasks
- Extract — Pull emails since last check, filter signal from noise
- Summarize — 1-3 bullet summary per important email
- Draft responses — Routine replies (confirmations, acknowledgments, scheduling) written at "ready to send" quality
- Flag — Anything requiring E-man's judgment, signature, or decision gets highlighted
Briefing Format
📬 Morning Briefing — [Date]
🔴 NEEDS ACTION (E-man's call)
• [Email subject] — 1 line what it is + recommended action
🟡 FYI (No action needed)
• [Email subject] — 1 line what it is
✅ ROUTINE — Drafted and ready to send
• [Email subject] → [Draft response ready]
Response Drafting Rules
- Keep under 3 sentences for routine replies
- Warm but professional
- If uncertain about tone → ask Ebi before sending
- Never send anything financial, legal, or personal without E-man reviewing
Tools
gws gmail list— fetch emailsgws gmail message— read full emailgws gmail reply— send draft (after Ebi reviews)- AgentMail SDK — for [email protected] if needed
Output
- Briefing delivered to Ebi (orchestrator) as markdown
- Ebi reviews, curates, delivers to E-man via Telegram
- Routine drafts flagged "ready to send" — Eman approves before sending
安全使用建议
Before installing, ask the publisher to enumerate exactly which credentials and OAuth scopes the skill needs (Gmail read/send scopes, AgentMail API key, etc.), and where data will be sent or stored (confirm [email protected] ownership and logging/retention). Do not grant mailbox access until you: 1) confirm the required env vars/credentials and minimal scopes, 2) verify the AgentMail endpoint and data forwarding rules, and 3) test the skill on a non-sensitive test mailbox. If the publisher cannot provide explicit credential requirements and a data-flow diagram, consider this skill too opaque to trust.
功能分析
Type: OpenClaw Skill
Name: briefing-agent
Version: 1.0.0
The skill bundle defines a 'Briefing Agent' for email management and summarization. The instructions in SKILL.md are well-defined and include explicit safety constraints, such as requiring human approval for sensitive topics (financial/legal) and ensuring drafts are reviewed before sending. The tools used (gws gmail and AgentMail SDK) are consistent with the stated purpose, and no indicators of malicious intent, data exfiltration, or unauthorized execution were found.
能力评估
Purpose & Capability
The skill's stated purpose is to read, summarize, and draft/reply to emails. The SKILL.md explicitly relies on Gmail commands (gws gmail list/message/reply) and an AgentMail SDK for [email protected], but the registry metadata lists no required credentials, primaryEnv, or config paths. Access to Gmail and AgentMail is required to perform the advertised tasks, so the declared requirements are incomplete and therefore incoherent with the claimed purpose.
Instruction Scope
Runtime instructions tell the agent to pull emails 'since last check', read full messages, and send replies via gws and AgentMail. The skill does not specify how authentication will occur, what mailbox(s) are targeted, retention/forwarding behavior, or limits on what is shared with the AgentMail address. That gap gives the agent broad discretion over reading and transmitting potentially sensitive email content.
Install Mechanism
This is instruction-only with no install spec or code files, so nothing is written to disk during install. That lowers supply-chain risk; however, the lack of install steps also means credential/config expectations are unspecified.
Credentials
No environment variables, credentials, or config paths are declared, yet the skill requires mailbox read/send capabilities and an AgentMail integration. Requesting no credentials while instructing access to email/third-party endpoints is disproportionate and opaque — users need to know exactly which tokens, scopes, and external endpoints will be used.
Persistence & Privilege
The skill is not marked 'always:true' and does not request persistent system-wide modifications. Autonomous invocation is allowed (platform default) but that alone is not an additional privilege here.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install briefing-agent - 安装完成后，直接呼叫该 Skill 的名称或使用
/briefing-agent触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.0
briefing-agent v1.0.0
- Initial release of the Briefing Agent skill.
- Reads and filters recent emails for actionable content.
- Summarizes important emails in concise bullet points.
- Drafts routine responses, keeping replies short and professional.
- Flags emails requiring E-man's decision or review.
- Outputs a clear, categorized morning briefing in markdown for supervisor review.
元数据
常见问题
Briefing Agent 是什么？
Reads emails, filters key info, summarizes important messages, drafts routine replies, and flags items needing your attention for a clear morning briefing. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 81 次。
如何安装 Briefing Agent？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install briefing-agent」即可一键安装，无需额外配置。
Briefing Agent 是免费的吗？
是的，Briefing Agent 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Briefing Agent 支持哪些平台？
Briefing Agent 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Briefing Agent？
由 EmanxChan（@emanxchan）开发并维护，当前版本 v1.0.0。
推荐 Skills