gutsofpratik/horizon-skills/source-04f4a95ae78cd336ebbd1326

Your Rootly Dashboard

You are building a personal status board for the current Rootly user. Read-only, focused, and quick to scan.

Workflow

1. Identify the user

Call mcp__rootly__getCurrentUser. Capture:

• user_id (UUID) — for filtering downstream calls
• name and email — for the header
• Team membership if present in the response

If getCurrentUser fails, stop with a clear auth error message.

2. Pull the three data slices in parallel

You may issue these tool calls in parallel where the model supports it. Keep each call narrow — we want a fast dashboard, not a full audit.

a. Open action items

• Call mcp__rootly__listAllIncidentActionItems (first page, page_size = 50).
• Filter to items where assigned_to_user_id (or equivalent field) matches the user, with status not in done/closed/cancelled.

b. Active incidents you're on

• Call mcp__rootly__listIncidents with filter_status=started and filter_user_id=<user_id> if supported, or with the user's team via filter_team_ids if individual filtering isn't available.
• Page size 25 is plenty.

c. Upcoming shifts

• Call mcp__rootly__listShifts for the current user, scoped to the next 7 days.
• If individual user filtering isn't available, fall back to mcp__rootly__list_shifts and filter client-side.

3. Render the dashboard

## [User name]'s Rootly Dashboard
*[email] — [primary team if known]*
 
### Active Incidents ([N])
[For each: severity badge, INC-XXXX, title, your role, started X ago]
- 🔴 [INC-XXXX] [title] — your role: [responder/IC/observer], started [duration] ago
- 🟡 [INC-YYYY] [title] — your role: [responder], started [duration] ago
 
[If zero: "No active incidents you're responding to."]
 
### Open Action Items ([N])
[Cluster by incident, max 10 items]
 
#### [INC-XXXX] [title]
- [ ] [description] — due [date or "no due date"]
- [ ] [description]
 
[If zero: "No open action items assigned to you."]
 
### Upcoming On-Call ([N] shifts in the next 7 days)
- [Schedule name]: [start] → [end] ([role/level if available])
- [Schedule name]: [start] → [end]
 
[If currently on call:]
**Currently on call**: [Schedule name] until [end time].
 
[If zero: "No on-call shifts in the next 7 days."]
 
---
 
### Quick links
- `/rootly:status` — service health overview
- `/rootly:oncall` — full team on-call schedule
- `/rootly:action add <incident> "<desc>"` — capture a follow-up
- `/rootly:swap` — request a shift swap

4. Tone & formatting

• Compact. This is meant to be glanceable, not a full report.
• Use severity emoji or symbols if they render in the user's terminal: 🔴 (critical), 🟠 (high), 🟡 (medium), 🟢 (low). If unsure, use plain text labels.
• Cap each section at the most-relevant 10 items. If there are more, append a single line: "(+N more — run /rootly:status or check Rootly UI)".

5. Error handling

• If any one slice fails (e.g. shifts API returns 500), render the other two and add a short note: "Could not load on-call shifts: [error]".
• Never leave a section blank with no explanation.