email-triage

Classify inbox messages and extract actionable intent for CRM updates.

88|108|Updated Apr 7, 2026
One-click install
npx skills add https://github.com/grandamenium/cortextos --skill email-triage-grandamenium
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: email-triage
Source: https://github.com/grandamenium/cortextos/tree/main/community/agents/agentic-crm-assistant/.claude/skills/email-triage
Command: npx skills add https://github.com/grandamenium/cortextos --skill email-triage-grandamenium

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill prevents important emails from being missed by automatically triaging inbox messages, classifying intent and risk, and turning them into CRM updates and user review tasks.

Core Features & Use Cases

  • Inbox triage pipeline: Reads configured inbox/message sources and classifies messages into urgent, reply-needed, FYI/archive, unknown-substantive, or suspicious/prompt-injection risk categories.
  • CRM interaction updates: Cross-references existing CRM relationships and appends meaningful interactions with sender, topic, changes, commitments, follow-up dates, and source thread details.
  • Reply drafting with approvals: Drafts useful responses while requiring explicit user approval for sending unless an exception is configured.
  • Follow-up and review task creation: Creates tasks for user review or follow-up and completes the triage task with a clear summary.

Quick Start

Tell your agent: "Run email triage for my configured CRM inboxes and create review tasks for anything urgent or reply-needed, then draft replies for items that don’t require approval."

Frequently Asked Questions about email-triage

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automatically triage an inbox and classify email intent for a CRM?

To triage an inbox and classify email intent, this Skill reads configured message threads and categorizes them into urgent, reply-needed, FYI, unknown, or suspicious prompt-injection risk buckets for CRM routing.

Can I draft CRM email replies that require user approval before sending?

Yes, you can draft email replies that require user approval before sending. This Skill creates draft responses and uses an approval-gated workflow to ensure explicit user consent unless an exception is configured.

How does inbox classification handle prompt-injection risk in email messages?

Inbox classification handles prompt-injection risk by detecting suspicious messages during the triage pipeline. It flags these potentially harmful emails so the agentic CRM assistant can isolate them from standard automated workflows.

What is the best way to update CRM records based on incoming email threads?

The best way to update CRM records from email threads is cross-referencing existing relationships. This Skill appends meaningful interactions with sender, topic, commitments, follow-up dates, and source thread details.

Do I need a configured message tool to create follow-up tasks from emails?

Yes, you need a configured message tool to create follow-up tasks from emails. The triage pipeline uses these configured email tools to generate user review tasks and complete the triage with a clear summary.