triage-messages

name	triage-messages
description	Batch-processes Slack messages and emails to surface what needs the manager's attention, categorised by urgency and type. Designed for batch-responder managers who do Slack sweeps rather than staying in reactive mode. Supports effective communication and responsiveness. Never drafts replies, only surfaces and prioritises.

Triage Messages

Principle: "You are responsible." This skill surfaces and categorises. The manager decides what to act on and how.

Scans unread Slack messages and recent emails, categorises them by urgency and type, and presents a prioritised overview.

When to Use

• Morning routine: catch up on what happened overnight
• After a block of focus time or meetings
• When the manager says "what did I miss?", "catch me up", "triage my messages"
• Can be scoped to a time window: "triage since yesterday", "triage last 2 hours"

Instructions

If any MCP connector is unavailable, follow the connector unavailability protocol in references/operating-principles.md.

1. Determine Time Window

Default: since the manager's last likely check-in (use calendar to estimate, if they were in back-to-back meetings for 3 hours, scan those 3 hours).

If specified: use the provided time window.

If unclear, ask:

What time period should I triage? Options:
- Since this morning
- Last [N] hours
- Since yesterday
- Custom range

2. Load Manager Context

Read from manager-context/ (if available):

• manager-profile.md: to know who the direct reports are (their messages get higher priority)
• terminology.md: to decode internal shorthand in messages
• team/: to understand team members' projects and responsibilities
• sources.md: to know which customer/project channels to monitor

3. Scan Slack

Search for messages in the time window:

Priority channels (scan first):

• Channels where the manager was @-mentioned
• DMs to the manager
• Team channels (from manager-context)
• Customer/project channels (from manager-context)

Broader channels (scan for relevance):

• Company-wide channels
• Cross-functional channels the manager is in

For each message/thread, capture:

• Who sent it
• Channel
• Content summary (1 line)
• Whether the manager was mentioned or it's in a direct thread
• Link to the message

4. Scan Email

Search Gmail for messages in the time window:

• Unread emails
• Emails where the manager is in To: (not just CC:)
• Emails from direct reports or known stakeholders
• Emails with urgent/action keywords

For each email, capture:

• Sender
• Subject
• Brief summary
• Whether it requires action or is FYI

5. Categorise

Sort everything into these buckets:

🔴 Needs Your Decision

• Blockers waiting on the manager
• Approval requests
• Escalations
• Time-sensitive decisions
• Criteria: someone is explicitly waiting for the manager's input

🟡 Team Member Needs

• Questions from direct reports
• Requests for help or guidance
• Friction signals (repeated blockers, frustration)
• Criteria: a team member needs support but isn't fully blocked

🔵 FYI / Context

• Status updates
• Announcements
• Cross-functional threads where the manager should be aware
• Criteria: useful to know, no action needed right now

🟢 Customer-Related (if applicable)

• Messages in customer/project channels
• Customer-related emails
• Delivery updates, escalations, timeline changes

6. Present the Triage

Read references/output-template.md for the full output template structure.

7. Offer Follow-Up

That's your triage for [time window]. Want me to:
- Dig deeper on any specific item?
- Prep context for responding to something?
- Run /plan-priorities to fold these into your day?

Important Notes

Read references/operating-principles.md for shared operating principles (data scope, DM flagging, signals vs diagnoses, connector unavailability).

Additional notes specific to this skill:

• Never draft replies. The manager writes their own messages. This skill only surfaces and categorises.
• Don't over-categorise as urgent. Reserve the red category for genuine blockers. Most things are FYI.
• Link everything. Every item should link directly to the source message/email.
• Decode shorthand. Use terminology.md to translate internal terms so the triage is immediately clear.
• If volume is high (50+ messages), summarise the FYI bucket rather than listing every item.