outlook-email-reply-drafting

$npx mdskill add openai/plugins/outlook-email-reply-drafting

Drafts Outlook email replies from mailbox context for safe review before sending.

  • Helps users reply to email threads, decide on reply-all, and prepare polished responses.
  • Integrates with Outlook via actions like search_messages, fetch_message, and create_reply_draft.
  • Decides by reading recent messages to understand tone and commitments before drafting.
  • Presents results as plain-text drafts unless explicitly instructed to send emails.

SKILL.md

.github/skills/outlook-email-reply-draftingView on GitHub ↗
---
name: outlook-email-reply-drafting
description: Draft Outlook email replies safely from connected mailbox context. Use when the user wants to reply to a thread, decide whether to reply-all, prepare a draft before sending, or turn the latest Outlook message into a polished response.
---

# Outlook Email Reply Drafting

## Overview

Use this skill when the reply itself is the task. Read enough mailbox context to understand the latest ask, then default to a draft unless the user clearly asked you to send.

Outlook reply writes are plain-text only. If the user asks for HTML formatting, styled sections, or layout-specific markup, say briefly that Outlook reply actions here only support plain text, then convert the request into a clean plain-text reply instead of planning an HTML body.

## Relevant Actions

- Use `search_messages` or `list_messages` to find the right thread when the user names a topic rather than a concrete message.
- Use `fetch_message` or `fetch_messages_batch` when the latest snippet is not enough to understand tone, commitments, or the actual ask.
- Use `create_reply_draft` for reply and reply-all drafts tied to an existing message.
- Use `reply_to_email` only when the user explicitly asked to send and the content is fully grounded.
- Use `draft_email` only for net-new messages or when the task is not a direct reply to an existing message.

## Workflow

1. Identify the exact source message or thread before drafting.
2. Read the most recent message first, then enough nearby context to understand participants, status, commitments, and tone.
3. Decide whether reply-all is necessary based on shared context, not just recipient count.
4. Draft the reply in the thread's tone unless the user asks for a deliberate change, but keep the draft plain text.
5. If the draft depends on missing facts, produce the best draft you can and list the unresolved details separately.
6. If the user later approves sending, reuse the thread-grounded draft instead of recreating the reply from scratch.

## Safety

- Preserve dates, commitments, names, links, and quoted facts unless the user asks to change them.
- Do not invent availability, approvals, ownership, or promises that are not already established in mailbox context.
- Treat reply-all as a deliberate choice. If the audience is ambiguous, explain the safest default.
- If the user says "send" but the content still depends on unstated choices, stop and ask the narrowest necessary confirmation question.
- Do not promise HTML email formatting or markup-specific rendering. Express structure with plain-text paragraphs, lists, and raw links, and tell the user when that means a requested formatting detail cannot be preserved exactly.

## Output

- Provide a ready-to-send plain-text draft with greeting, body, and closing when appropriate.
- If important assumptions remain, list them immediately after the draft.
- If you are recommending a reply-all decision, say why in one short line.

More from openai/plugins

SkillDescription
accessibility-and-inclusive-visualizationMake data visualizations accessible and inclusive. Use when the user needs chart or diagram accessibility guidance, text alternatives for complex visuals, color and contrast review, keyboard support, reduced-motion behavior for animation or parallax, or an accessibility QA workflow for exported figures, UML-like diagrams, and dashboards.
agent-browserBrowser automation CLI for AI agents. Use when the user needs to interact with websites, verify dev server output, test web apps, navigate pages, fill forms, click buttons, take screenshots, extract data, or automate any browser task. Also triggers when a dev server starts so you can verify it visually.
agent-browser-verifyAutomated browser verification for dev servers. Triggers when a dev server starts to run a visual gut-check with agent-browser — verifies the page loads, checks for console errors, validates key UI elements, and reports pass/fail before continuing.
agents-sdkBuild AI agents on Cloudflare Workers using the Agents SDK. Load when creating stateful agents, durable workflows, real-time WebSocket apps, scheduled tasks, MCP servers, or chat applications. Covers Agent class, state management, callable RPC, Workflows integration, and React hooks. Biases towards retrieval from Cloudflare docs over pre-trained knowledge.
ai-elementsAI Elements component library guidance — pre-built React components for AI interfaces built on shadcn/ui. Use when building chat UIs, message displays, tool call rendering, streaming responses, reasoning panels, or any AI-native interface with the AI SDK.
ai-gatewayVercel AI Gateway expert guidance. Use when configuring model routing, provider failover, cost tracking, or managing multiple AI providers through a unified API.
ai-generation-persistenceAI generation persistence patterns — unique IDs, addressable URLs, database storage, and cost tracking for every LLM generation
ai-sdkVercel AI SDK expert guidance. Use when building AI-powered features — chat interfaces, text generation, structured output, tool calling, agents, MCP integration, streaming, embeddings, reranking, image generation, or working with any LLM provider.
aiq-deploy|
aiq-research|