--- name: oncall-rota-builder description: Build a fair, documented on-call rota for a small IT team — with coverage pattern, primary/secondary, handoff protocol, compensation, bank-holiday rules, exception handling, and a published schedule. version: 1.0.0 author: VantagePoint Networks author_url: https://www.vpnetworks.co.uk audience: IT Managers, Service Delivery Managers, MSP Operations Leads, SRE Managers, Solo Founders about to hire their second engineer output_format: Formatted Markdown rota document with coverage pattern, rota table, handoff protocol, compensation model, exception procedures, and escalation path. license: MIT last-reviewed: 2026-04 --- # On-Call Rota Builder A Claude Code skill for designing the on-call rota that every growing IT team eventually needs — before burnout drives people out, or before a single-person dependency creates a real crisis. ## How to use this skill 1. Download this `SKILL.md` file. 2. Place it in `~/.claude/commands/` (macOS/Linux) or `%USERPROFILE%\.claude\commands\` (Windows). 3. In Claude Code, run `/oncall-rota-builder`. Describe the team and service. Answer the clarifying questions. Receive the rota document. ## When to use this - You've been the informal on-call for everyone and it's unsustainable. - You hired your second or third engineer and need a structure. - An MSP is onboarding a new client and the contract requires 24/7 coverage. - A tribunal, insurer, or client asked "what's your out-of-hours support posture?" - You've had a close call where nobody was reachable during an incident. ## What you'll get A single Markdown document containing: - **Coverage pattern** (who's on when, for how long) - **Rota table** (named weeks ahead + holiday handling) - **Primary / secondary / escalation** roles - **Handoff protocol** (at start / end of shift) - **Compensation model** (time off, allowance, overtime) - **Bank-holiday and weekend rules** - **Exception handling** (sickness, personal emergency, swap requests) - **Escalation path** (when primary can't cope) - **Rules for the on-call engineer** (response time, tooling, dress-for-work-from-home) - **Review cadence** (monthly rota health-check) ## Clarifying questions I will ask you 1. **Team size (engineers available for rota)?** 2. **Coverage window required?** (24/7 / business hours extended / evenings + weekends only) 3. **Geographic spread?** (all UK / follow-the-sun with EU / US offices) 4. **Services in scope?** (which need immediate response vs. next-business-day) 5. **Expected alert volume per week?** (light, moderate, heavy) 6. **Existing tooling?** (PagerDuty, Opsgenie, phone-tree, nothing) 7. **Legal / contractual response-time commitments?** 8. **Budget for on-call compensation?** (allowance / hourly / TOIL / none) 9. **Any staff who can't be on call?** (caring responsibilities, medical, contractual) 10. **Bank-holiday expectation?** (full coverage / reduced / skip) 11. **Swap flexibility?** (strict rota / swaps allowed / tradeable) 12. **What's broken today?** (informal-only / one person carries it all / coverage gaps / response-time misses) ## Output template ```markdown # On-Call Rota — — **Owner:** · **Tool:** · **Review:** monthly ## 1. Coverage Model - **Coverage window:** - **Services in scope:** - **Response targets:** - Acknowledgement: < 15 min (P1), < 30 min (P2) - Initial diagnosis / triage: < 1 hour - Engagement of secondary if needed: < 30 min from primary request - **Out of scope:** ## 2. Roles | Role | Responsibility | |---|---| | Primary on-call | First responder. Carries pager. Must be reachable and capable of acting within SLA. | | Secondary on-call | Backup. Consulted when primary needs help or is unavailable. Also carries pager, quieter. | | Escalation (Manager-on-call) | Called when primary + secondary both cannot resolve, or when business/customer decisions needed. | | Subject-matter experts (SMEs) | Non-rota specialists who may be paged for specific issues (DBA, Network SME, Security). Informal/best-effort. | ## 3. Rota Table (next 12 weeks, indicative) | Week | Primary | Secondary | Manager-on-call | |---|---|---|---| | W1 | Eng A | Eng B | Mgr | | W2 | Eng B | Eng C | Mgr | | W3 | Eng C | Eng D | Mgr | | W4 | Eng D | Eng A | Mgr | | W5 | Eng A | Eng B | Mgr | | ... | | | | **Rotation pattern:** weekly (Mon 09:00 to Mon 09:00 handoff). **Maximum frequency:** 1-in-4 (never more than every 4th week for primary). **Maximum consecutive weeks:** 1 on primary, optional 1 on secondary immediately after. ## 4. Handoff Protocol Every shift ends and begins with a handoff. Template: see `/oncall-handoff-writer` skill output. Handoff meeting: - **Monday 09:00:** 15-min call between outgoing and incoming primary. - Outgoing presents: open issues, recent changes, watch items, any silenced alerts, any scheduled changes. - Incoming confirms: has pager access, runbook locations, contact list, no conflicts for the week. ## 5. Compensation Model | Component | Value | |---|---| | On-call allowance (per week of primary) | £ | | On-call allowance (per week of secondary) | £ | | Paid call-out (actual work during out-of-hours) | hourly rate × 1.5, minimum hours | | TOIL (time off in lieu) — alternative | 1:1 hours for work done between 22:00 - 08:00 | | Bank holiday on-call | double allowance | | Shift swap approved | no financial adjustment; hours balance managed by rota owner | ## 6. Bank Holidays & Weekends - Full coverage on all working days. - Reduced-alert weekends (P1 only, P2 → next working day unless customer SLA differs). - Bank-holiday weeks covered by the primary whose normal rotation hits that week. - December / summer holidays: named swap-partners agreed 8 weeks in advance. ## 7. Rules for the Primary On-Call - **Acknowledge every page** within target even if you're about to investigate — silence is bad signal. - **Reachable means:** phone on, volume up, in a location with signal and charging available. - **Work-from-home-capable:** laptop + power + network within reach. - **If you consume alcohol or anything impairing, hand off to secondary immediately.** Not punishable; expected. - **You don't have to solve every issue alone.** Escalate to secondary, SMEs, or escalation-manager. - **Document everything in the incident channel.** Even if unsure. Audit trail matters. - **End of shift, you hand off.** You don't keep carrying an open issue into your personal time. ## 8. Exception Handling ### Sickness during shift - Notify rota owner + secondary immediately. - Secondary covers until rota owner finds replacement or secondary becomes primary. - Full compensation for the replacement / secondary taking over. ### Personal emergency - Phone / text rota owner. No need for details. - Same handoff as sickness. ### Swap request - Both parties email rota owner ≥ 48 hours before (where possible). - Rota owner confirms in writing. - Swap registered in the tool and updated in the schedule. ### Opt-out - Documented reasons: caring responsibilities, medical, contractual (e.g. part-time). - Rota owner ensures opt-outs don't concentrate load on remaining staff; hiring discussion if it does. ## 9. Escalation Path When primary is overwhelmed or stuck: 1. Secondary (already warm via handoff) 2. SME (if technical specific) 3. Manager-on-call (decisions, not technical) 4. External vendor TAM / MSP escalation line 5. Client sponsor (only for customer-sponsored incidents, typically P1+) Each escalation step is documented in the incident channel so the trail is audit-ready. ## 10. Review Cadence ### Weekly - Rota owner reviews: pages received, who was paged, any SLA misses, any swaps. ### Monthly (30-min meeting) - Total pages / person. - Distribution (fair?). - Issues with handoffs. - Swap frequency (high = pattern?). - Upcoming holidays / gaps. - Changes to rota or policy agreed here. ### Quarterly - Compensation review (still appropriate for the load?). - Tooling review. - Team satisfaction sentiment. ### Annual - Full rota policy review. - Hiring needs based on load trends. ## 11. Tooling - **Pager:** . - **Schedule management:** integrated with pager tool. - **Handoff document:** stored in , shared with incoming primary. - **Incident documentation:** . - **Runbook library:** . ## 12. What Good Looks Like - No engineer on call more than 1-in-4 weeks. - Fewer than pages per primary-week on average; fewer than of those actionable after tuning (link to alert-triage-reviewer). - Handoff meeting happens every Monday. - Zero SLA misses in the quarter, or documented reasons. - Team retention on the rota > 12 months. ## 13. What Bad Looks Like (and what to do) - Same person carrying > 40% of pages → distribute via tuning or hiring. - Handoff meetings skipped → rota owner enforces; it's not optional. - Swap frequency > 50% of weeks → life-reasons or rota design problem — investigate. - Regular out-of-hours work without compensation tracking → audit TOIL usage; enforce logging. - People leaving citing on-call → urgent review: too often, too noisy, or not enough support. ``` ## Example invocation **User:** "5-engineer IT team at a 120-person London fintech. 24/7 coverage needed (FCA operational resilience). Currently: informal phone-tree, senior engineer carries most of it, people burning out." **What the skill will do:** 1. Ask 12 questions, pressing on: who's currently opting out and why, the FCA resilience impact-tolerance for their IBSes (informs acknowledgement SLA), whether any engineers can't do on-call contractually. 2. Produce the rota with: - 1-in-4 primary rotation (with 5 engineers, one week on, three off) - Paired secondary (different engineer, quieter week) - Senior-engineer carrying manager-on-call role (not primary) — removes the burden they've had - Allowance model: £250/week primary, £100/week secondary, double for bank holidays - Specific Monday 09:00 handoff meeting protocol - 24/7 coverage commitments matched to FCA-level SLAs 3. Flag that hiring a 6th engineer within 6-9 months is likely necessary given the coverage + growth pattern. ## Notes for the requester - **1-in-4 is the healthy maximum for sustained primary rotation.** More than that and attrition accelerates. - **Never just "secondary is the same person who's also the backup DBA".** The rota must be distinct people, even if they're informally consulted as SMEs. - **Compensation is not a bonus — it's payment for restricted time.** Treat it as a legal obligation, documented. - **Rota fairness is a retention issue.** Pages-per-engineer over a quarter should be within ~20% of the mean. Outliers indicate the rota or alert hygiene needs work. - **Handoff meeting is non-negotiable.** Skipping it leads to "wait, what was that alert from last Wednesday?" during an incident. - **Good looks like:** engineers stay on the rota willingly, swaps happen for life-reasons not rota-reasons, and the rota scales without heroics when the team grows or shrinks. --- *VantagePoint Networks · · Authored by Hak · Free under the MIT licence*