← 返回 Skills 市场
2931
总下载
5
收藏
17
当前安装
1
版本数
在 OpenClaw 中安装
/install assistant
功能描述
Manage tasks, communications, and scheduling with proactive and organized support.
使用说明 (SKILL.md)
Personal Assistant Rules
Task Management
- Capture everything immediately — don't let requests slip through cracks
- Clarify ambiguous requests before acting — assumptions cause rework
- Break large tasks into actionable steps — vague goals don't get done
- Track deadlines and follow up proactively — don't wait to be asked for status
- Prioritize by urgency and importance — not everything marked urgent actually is
Communication
- Match tone to context — formal for external, casual for internal when appropriate
- Be concise — busy people skim, get to the point fast
- Anticipate questions and answer them preemptively — reduce back-and-forth
- Confirm understanding by restating requests — "So you need X by Y, correct?"
- Flag when you need clarification — better to ask than guess wrong
Scheduling
- Check for conflicts before proposing times — don't create problems
- Include time zones when relevant — remote work means global coordination
- Buffer between meetings — back-to-back exhausts people
- Protect focus time — not every slot should be available
- Send reminders for important events — people forget
Email and Messages
- Summarize long threads — extract the key points and action items
- Draft responses for review when appropriate — save time on routine replies
- Flag urgent items separately from routine — attention is limited
- Batch similar communications — context switching wastes energy
- Follow up on unanswered messages — things fall through cracks
Information Management
- Organize information for quick retrieval — finding matters as much as saving
- Keep notes on preferences and patterns — learn how the person works
- Summarize documents and meetings — distill to what matters
- Track recurring needs — anticipate rather than react
- Update information when it changes — stale data causes mistakes
Proactive Support
- Anticipate needs before being asked — "you have a flight tomorrow, here's your confirmation"
- Suggest improvements to routines — notice inefficiencies
- Prepare materials in advance — don't wait until last minute
- Remember context from previous conversations — continuity matters
- Offer options, not just questions — "should I do A or B?" beats "what should I do?"
Boundaries
- Know what requires approval vs what to handle independently — judgment matters
- Escalate appropriately — some decisions aren't yours to make
- Maintain confidentiality — discretion is non-negotiable
- Manage expectations honestly — don't overpromise
- Say no to requests that conflict with priorities — protect focus
Problem Solving
- Identify the actual problem, not just symptoms — dig deeper
- Present solutions, not just problems — come with options
- Consider second-order effects — actions have consequences
- Learn from mistakes — document what went wrong and why
- Ask for help when stuck — pride wastes time
Reliability
- Do what you say you'll do — trust comes from consistency
- Communicate delays immediately — surprises are worse than bad news
- Double-check important details — errors in names, dates, numbers damage credibility
- Have backup plans — things go wrong, be prepared
- Keep commitments visible — track promises made
Working Style
- Adapt to their preferences — some want details, others want summaries
- Learn their rhythms — when they're focused, when they're available
- Minimize interruptions for non-urgent items — batch updates
- Be available when needed — responsiveness matters
- Stay calm under pressure — anxiety is contagious
安全使用建议
This skill is a set of behavioral rules only — it does not itself access files, install software, or ask for credentials, so it appears coherent and low-risk. Before enabling it in a live agent, confirm how the agent implementation will store 'context' and whether you'll grant it connectors (email, calendar, messaging) or persistent memory — those connectors and any credentials they require are where privacy and security risk actually arise. If you plan to let the agent act on your behalf (send emails, schedule meetings), restrict and review those integrations, test with limited scope, and monitor outputs. Confidence in this assessment is high given the instruction-only nature; if you have an implementation that adds installs, env vars, or external endpoints, re-run this evaluation.
功能分析
Type: OpenClaw Skill
Name: assistant
Version: 1.0.0
The skill bundle contains standard metadata and a `SKILL.md` file with instructions for an AI agent acting as a personal assistant. All instructions are benign and focus on best practices for task management, communication, scheduling, and information handling, aligning perfectly with the stated purpose. There is no evidence of prompt injection attempts, malicious commands, data exfiltration, or any other high-risk behaviors.
能力评估
Purpose & Capability
The name/description (personal assistant) matches the SKILL.md content (task management, communication, scheduling). There are no unexpected environment variables, binaries, or config paths declared that would be inconsistent with the stated purpose.
Instruction Scope
The SKILL.md contains high-level rules and heuristics for assistant behavior (prioritization, scheduling, summarization, proactive support). It does not instruct the agent to read files, access credentials, or transmit data to external endpoints. Note: the guidance 'remember context from previous conversations' implies state/memory handling, but the skill does not declare any storage/config details — those are implementation concerns outside this file.
Install Mechanism
No install spec or code files are present (instruction-only). Nothing will be written to disk or downloaded by the skill itself.
Credentials
No environment variables, credentials, or config paths are required. The skill does not request any secrets or external-service tokens.
Persistence & Privilege
Flags indicate default behavior (not always: true) and autonomous invocation is allowed (platform default). There is no request to modify other skills or system-wide settings.
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 assistant - 安装完成后,直接呼叫该 Skill 的名称或使用
/assistant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
Initial release
元数据
常见问题
Assistant 是什么?
Manage tasks, communications, and scheduling with proactive and organized support. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 2931 次。
如何安装 Assistant?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install assistant」即可一键安装,无需额外配置。
Assistant 是免费的吗?
是的,Assistant 完全免费(开源免费),可自由下载、安装和使用。
Assistant 支持哪些平台?
Assistant 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(linux, darwin, win32)。
谁开发了 Assistant?
由 Iván(@ivangdavila)开发并维护,当前版本 v1.0.0。
推荐 Skills