| name | change-request |
| description | ITIL change request and CAB review workflow. This skill should be used when the user asks to "create a change request", "write an RFC", "plan a change", "CAB review", "change management", "schedule maintenance", "server migration plan", "upgrade plan", "AI model deployment", "GitOps change", or describes any planned modification to IT infrastructure or services. Produces a complete RFC document following ITIL v5 Change Enablement. |
| allowed-tools | ["Read","Write","Edit","Glob","Grep","Bash","WebSearch","WebFetch","AskUserQuestion"] |
| argument-hint | <description of the planned change> |
| user-invocable | true |
Change Request Workflow
Structured change management following ITIL 4 Change Enablement. Produces a complete RFC (Request for Change) document that serves as both the change plan and the CAB review submission.
Related Skills (auto-loaded)
- itil — change types (Standard/Normal/Emergency), CAB process, risk assessment matrix
- done-ops — verification gate before finalizing the change
Output
File: {OUTPUT_DIR}/change-request-{YYYY-MM-DD}-{slug}.md
- slug: lowercase kebab-case summary, e.g.
exchange-upgrade, ad-entra-migration
- Change ID (
CHG-{YYYYMMDD}-{NNN}): {NNN} is a daily sequence number starting at 001. If the organization uses a ticketing system, use that system's ID instead.
Default output directory: ./changes/. If a docs/changes/ directory exists in the project, use that instead. Ask the user if neither exists. A user-specified path takes precedence over all defaults.
All change request documents are written in English. Before writing to file, preview the content for the user in their preferred language.
Workflow
Step 1: Gather Change Details
Collect the following from the user (use AskUserQuestion for missing items):
| Field | Required | Example |
|---|
| What is being changed? | Yes | "Upgrade Exchange Server 2019 to 2022" |
| Why is this change needed? | Yes | "End of support, security compliance" |
| Which systems/services affected? | Yes | "Email service, Active Directory" |
| Expected downtime? | Yes | "4-hour maintenance window" |
| Proposed schedule? | Yes | "Saturday 02:00-06:00" |
| Who will implement? | Yes | "IT Infrastructure team" |
Step 2: Classify the Change
Determine the change type using the itil skill:
| Question | Standard | Normal | Emergency |
|---|
| Is this a routine, pre-approved procedure? | Yes → Standard | No | No |
| Is there a documented SOP for this? | Yes → Standard | Maybe | No |
| Is this time-critical (active incident)? | No | No | Yes → Emergency |
| Otherwise | — | → Normal | — |
Clarification: Compliance deadlines or business mandates with hard dates are Normal changes (with appropriate urgency), not Emergency. Emergency is reserved for active incidents or immediate security threats requiring same-day implementation.
AI / ML System Changes: If the change involves any AI or ML component (model update, prompt change, training data, RAG index, AI autonomy scope), load itil/references/change-enablement.md and follow the "AI / ML System Changes" section. Changes involving Cognition or Coordination capabilities require Normal (Significant) classification at minimum.
GitOps / IaC Changes: Changes delivered via pull request and automated pipeline may qualify as Standard if they meet pre-authorization criteria. See itil/references/change-enablement.md → "GitOps / IaC Change Pattern".
For Normal changes, further classify risk level:
- Minor: Low risk, Change Manager approval sufficient
- Significant/Major: Requires full CAB review
Step 3: Risk Assessment
Rate each dimension 1-3, multiply for Risk Score:
Impact (if change fails):
- 3 = High: critical service outage, data loss risk
- 2 = Medium: degraded service, limited users affected
- 1 = Low: minimal/no user impact
Likelihood (of failure):
- 3 = High: untested procedure, complex dependencies
- 2 = Medium: tested in non-prod, moderate complexity
- 1 = Low: routine procedure, well-documented, previously successful
| Risk Score | Level | Authorization |
|---|
| 7-9 | High | Full CAB + management approval |
| 4-6 | Medium | CAB review recommended |
| 1-3 | Low | Change Manager approval |
MUST include the risk assessment table in the RFC with this exact format.
Sustainability note: If the change involves significant compute (AI model deployment, server provisioning >10 units, cloud region migration), add the optional Sustainability assessment from itil/references/sustainability.md to the Risk Assessment section.
| Dimension | Rating | Justification |
|---|---|---|
| Impact (if failure) | {1-3} | {Reason} |
| Likelihood (of failure) | {1-3} | {Reason} |
| **Risk Score** | **{N}** | **{Low/Medium/High}** |
Present the assessment to the user for validation.
Step 4: Build the RFC Document
Create the change request file. MUST follow this exact structure — do not rearrange or rename sections:
# Request for Change: {TITLE}
| Field | Value |
|---|---|
| Change ID | CHG-{YYYYMMDD}-{NNN} |
| Change Type | Standard / Normal / Emergency |
| Risk Level | {Low / Medium / High} (Score: {N}) |
| Status | Draft |
| Requester | {NAME} |
| Date Submitted | {YYYY-MM-DD} |
| Proposed Window | {DATE TIME} to {DATE TIME} ({TIMEZONE}) |
## Business Justification
{Why this change is needed. What problem it solves. What happens if not implemented.}
## Scope and Impact
### Systems Affected
{List of systems, services, and configuration items.}
### Users Affected
{Who will experience downtime or changes. Estimated number.}
### Dependencies
{Other systems, teams, or changes this depends on.}
## Risk Assessment
| Dimension | Rating | Justification |
|---|---|---|
| Impact (if failure) | {1-3} | {Reason} |
| Likelihood (of failure) | {1-3} | {Reason} |
| **Risk Score** | **{N}** | **{Low/Medium/High}** |
### Risk Mitigation
{Steps to reduce risk — e.g., tested in staging, phased rollout.}
## Implementation Plan
### Pre-Implementation
| Step | Action | Owner | Est. Duration |
|---|---|---|---|
| 1 | {Pre-check or preparation} | {Name} | {Time} |
### Implementation
| Step | Action | Owner | Est. Duration |
|---|---|---|---|
| 1 | {Step description} | {Name} | {Time} |
### Post-Implementation
| Step | Action | Owner | Est. Duration |
|---|---|---|---|
| 1 | {Verification step} | {Name} | {Time} |
## Test / Validation Plan
| Check | Expected Result | Method |
|---|---|---|
| {What to verify} | {Expected outcome} | {How to check} |
## Rollback Plan
**Trigger criteria:** {When to initiate rollback — specific, measurable conditions.}
**Estimated rollback time:** {Duration}
| Step | Action | Owner | Est. Duration |
|---|---|---|---|
| 1 | {Rollback step} | {Name} | {Time} |
## Communication Plan
| When | Audience | Channel | Message |
|---|---|---|---|
| Before change | {Who} | {Email/Teams/Statuspage} | Maintenance notification |
| During (if issues) | {Who} | {Channel} | Status update |
| After completion | {Who} | {Channel} | Completion confirmation |
## Approvals
| Role | Name | Decision | Date |
|---|---|---|---|
| Change Manager | | | |
| CAB | | | |
| Technical Lead | | | |
## Related Records
{Linked incidents, problems, or previous changes.}
Step 5: CAB Review Preparation
For Normal (Major) and Emergency changes, prepare the CAB review summary:
- Load the CAB Review Checklist from
itil skill (references/change-enablement.md)
- Self-evaluate the RFC against each checklist item
- Flag any items that are incomplete or need discussion
- Present the checklist evaluation to the user
Step 6: Finalize
- Load done-ops verification gate
- Verify all required sections are complete:
- Business justification is clear
- Implementation plan has specific steps with owners
- Rollback plan has trigger criteria and tested procedure
- Communication plan identifies all audiences
- Risk assessment is realistic
- Present the final RFC for user review
- Note the authorization path: who needs to approve this change
Gotchas
- The RFC is the single document for the entire change lifecycle — plan, approval, implementation, and review all reference this document
- Phased changes (e.g., multi-weekend migrations): use separate Implementation Plan tables per phase, add phase gate criteria (go/no-go between phases), and note all proposed windows in the Schedule section
- Standard changes do not need individual CAB review, but the standard change catalog should be referenced
- Emergency changes still need an abbreviated RFC before implementation, with a full RFC completed within 48 hours
- Always ask about blackout windows — some organizations have change freezes around releases, holidays, or fiscal close
- Rollback plan must have specific trigger criteria, not just "if something goes wrong"
- If the change was triggered by an incident, link the incident ID
- Post-Implementation Review (PIR) is mandatory for Emergency changes (48h) and recommended for all Normal changes (5 business days)