@raishin/vanguard-frontier-agentic 2.9.0 → 2.11.0

package/skills/microsoft/d365-customer-insights-journeys/metadata.json

@@ -0,0 +1,32 @@
+{
+  "id": "d365-customer-insights-journeys",
+  "name": "D365 Customer Insights — Data & Journeys",
+  "type": "skill",
+  "provider": "microsoft",
+  "harnesses": [
+    "codex",
+    "claude-code",
+    "cursor",
+    "gemini",
+    "kiro",
+    "other"
+  ],
+  "summary": "Review Dynamics 365 Customer Insights — Data (CDP: data unification, segments, measures) and Customer Insights — Journeys (real-time marketing journeys, triggers, consent/compliance, channel orchestration) design and configuration. Enforces unified profile completeness, segment quality gates, consent model correctness, journey logic review, and compliance posture. Detects fragmented customer data, weak segmentation, broken consent handling, and low-ROI journey design. Refuses to approve production journey publish, bulk outreach, or consent-model changes without evidence of consent compliance and journey validation. Live-guard gated for production journey publish, consent-model changes, and segment-based bulk outreach.",
+  "source_type": "original",
+  "official_docs": [
+    "https://learn.microsoft.com/dynamics365/customer-insights/data/data-unification",
+    "https://learn.microsoft.com/dynamics365/customer-insights/data/segments",
+    "https://learn.microsoft.com/dynamics365/customer-insights/journeys/real-time-marketing-compliance-settings",
+    "https://learn.microsoft.com/dynamics365/customer-insights/journeys/real-time-marketing-email-text-consent",
+    "https://learn.microsoft.com/dynamics365/customer-insights/journeys/real-time-marketing-double-opt-in",
+    "https://learn.microsoft.com/dynamics365/customer-insights/journeys/ci-get-started",
+    "https://learn.microsoft.com/dynamics365/customer-insights/data/get-started"
+  ],
+  "security_notes": "Never approve production journey publish or bulk outreach without documented evidence of consent compliance review, audience segment validation, and journey logic sign-off. Production journey publish, consent-model changes, and segment-based bulk outreach are live-guard gated and must be escalated to the marketing operations lead and compliance owner before execution. Consent model changes — including compliance profile edits, purpose reconfiguration, and topic restructuring — must be reviewed for regulatory impact (GDPR, CAN-SPAM, CASL) before deployment. Do not accept segment membership count alone as evidence of segment quality; require rule review and profile source validation. Do not ask for credentials, tenant IDs, environment URLs, API keys, customer PII, or consent data exports. Treat every unvalidated consent migration, missing double opt-in configuration, and untested journey branch as a production risk requiring explicit human sign-off.",
+  "last_verified": "2026-06-17",
+  "path": "skills/microsoft/d365-customer-insights-journeys",
+  "author": "github: Raishin",
+  "version": "0.1.0",
+  "category": "data",
+  "companion_agents": ["d365-customer-insights-journeys-agent"]
+}

package/skills/microsoft/d365-customer-insights-journeys/references/official-sources.md

@@ -0,0 +1,46 @@
+# Official sources
+
+Use this reference only when you need source grounding for Dynamics 365 Customer Insights — Data or Customer Insights — Journeys behavior, data unification, segment design, consent compliance, or journey orchestration guidance.
+
+## Microsoft Learn documentation
+
+Use these as starting points, not as proof of the user's live environment configuration or consent posture:
+
+- https://learn.microsoft.com/dynamics365/customer-insights/data/data-unification
+- https://learn.microsoft.com/dynamics365/customer-insights/data/segments
+- https://learn.microsoft.com/dynamics365/customer-insights/data/measures
+- https://learn.microsoft.com/dynamics365/customer-insights/data/get-started
+- https://learn.microsoft.com/dynamics365/customer-insights/journeys/real-time-marketing-compliance-settings
+- https://learn.microsoft.com/dynamics365/customer-insights/journeys/real-time-marketing-email-text-consent
+- https://learn.microsoft.com/dynamics365/customer-insights/journeys/real-time-marketing-double-opt-in
+- https://learn.microsoft.com/dynamics365/customer-insights/journeys/ci-get-started
+- https://learn.microsoft.com/dynamics365/customer-insights/journeys/real-time-marketing-migrate-consent
+- https://learn.microsoft.com/dynamics365/customer-insights/journeys/unified-profile-segment-creation
+
+## Grounding rule
+
+Official documentation explains Dynamics 365 Customer Insights behavior, data unification mechanics, segment evaluation rules, and consent enforcement logic. It does not prove the user's actual unified profile completeness, segment membership accuracy, consent record state, or journey branch coverage. Prefer documented artifacts (unification run logs, segment inspection results, consent audit exports, journey test run evidence, compliance profile configuration screenshots) over inference.
+
+## Service facts (verified 2026-06-17)
+
+Customer Insights — Data (CDP) structure:
+- Customer Insights — Data is a customer data platform (CDP) that ingests data from multiple sources, runs identity resolution through deduplication and match rules, and produces unified customer profiles stored in Dataverse.
+- **Data sources**: ingested via Power Query connectors, Azure Data Lake Storage, Dataverse, or other supported connectors. Each source table must be mapped to semantic fields before unification.
+- **Unification phases**: source field mapping → deduplication → match rules → merge policies. All four phases must be configured and run before unified profiles are available.
+- **Segments**: evaluated against unified customer profiles. Segment rules can reference profile attributes, activities, measures, and enrichments. Segments must be refreshed (scheduled or on-demand) before they reflect current profile data.
+- **Measures**: calculated KPIs based on unified profile data. Measures can be used as segment criteria in Customer Insights — Journeys.
+
+Consent model (Customer Insights — Journeys):
+- Consent is captured and enforced at the **contact point level** (email address, phone number, or custom channel address), not at the contact/lead record level.
+- The consent hierarchy: compliance profile → purpose → topic. A compliance profile must be configured before journeys can enforce consent.
+- **Double opt-in**: must be explicitly enabled per compliance profile. Enabling the feature switch does not automatically apply double opt-in to all forms; additional configuration per compliance profile is required.
+- The `DoNotEmail`, `DoNotBulkEmail`, and `DoNotTrack` fields on the contact record are **not** evaluated by default in real-time journeys. To enforce them, the "Check contact consent in real-time journeys" feature switch must be enabled.
+- Consent migration from outbound marketing or external systems requires manual import or the Load Consent function; contact point consent records are not automatically created from legacy field values.
+
+Journey orchestration:
+- Real-time journeys support segment-based entry (scheduled) and trigger-based entry (event-driven).
+- Journey branches, wait conditions, and channel steps must be tested before production publish. A journey cannot be edited after it is set to Live status without creating a new version.
+- Production journey publish is irreversible without stopping the journey; stopping a live journey affects all in-progress participants.
+
+Certification anchor:
+- MB-260 (Customer Insights Data Specialist) was retired 2026-06-17 per latest available information; MB-220 (Marketing) was also retired. Verify current certification and Applied Skills offerings on Microsoft Learn before citing exam references. (E4: verify before citing.)

package/skills/microsoft/d365-customer-insights-journeys/references/safety-checklist.md

@@ -0,0 +1,41 @@
+# Safety checklist
+
+Use this reference before any recommendation involving production journey publish authorization, consent-model changes, bulk outreach approval, or compliance-impacting configuration changes in Dynamics 365 Customer Insights — Data or Customer Insights — Journeys.
+
+## Non-negotiables
+
+- Never ask users to paste credentials, API keys, tenant IDs, environment URLs, client secrets, customer PII, or consent data exports into chat.
+- Use documented artifacts or sanitized user-provided evidence for live-state claims; otherwise use documentation and label the evidence level.
+- Do not invent unified profile counts, segment membership sizes, consent record states, journey test results, or compliance profile configurations.
+- Require explicit human approval before recommending production journey publish, consent-model changes, or bulk outreach execution.
+- Use current official Microsoft Learn documentation for Dynamics 365 Customer Insights — Data and Customer Insights — Journeys behavior.
+- Keep recommendations least-change, reversible, and scoped to the workload in question.
+- Production journey publish, consent-model changes, and segment-based bulk outreach are live-guard gated. Always escalate to the marketing operations lead and named compliance owner before execution.
+
+## Stress checks
+
+- Have all data sources been ingested and mapped before unification runs were executed?
+- Has the unified profile configuration been validated with sample profile results showing correct identity resolution?
+- Do segment rules reference validated profile attributes with a refresh schedule aligned to journey entry cadence?
+- Has at least one compliance profile been configured with purpose and topic hierarchy before any journey is published?
+- Is double opt-in configured on all compliance profiles where required by applicable regulation (GDPR, CAN-SPAM, CASL)?
+- Has legacy consent data been fully migrated from outbound marketing or external systems before bulk outreach begins?
+- Have all journey branches — including suppression, exit, error, and no-match paths — been tested in a non-production environment?
+- Has the compliance owner signed off on the consent model and unsubscribe path?
+- Is there a post-launch monitoring plan for engagement metrics, consent opt-out rates, and journey error rates?
+
+## Evidence labels
+
+Use `live evidence`, `documented artifact`, `user-provided evidence`, `documentation-based`, or `inference`. Documentation alone never proves the user's actual unified profile state, segment membership accuracy, consent record coverage, or journey branch coverage.
+
+## Live-guard gate
+
+The following actions require explicit human confirmation and are out of scope for automated execution:
+
+- Publishing a journey to Live status in the production environment
+- Modifying compliance profiles, purposes, or topics in the production consent model
+- Executing segment-based bulk email, SMS, or push notification outreach in production
+- Migrating or bulk-loading consent records into the production consent center
+- Enabling or disabling double opt-in on a compliance profile in production
+- Authorizing suppression list changes that affect production outreach
+- Signing off on regulatory compliance posture on behalf of the compliance owner

package/skills/microsoft/d365-customer-insights-journeys/references/workflow-and-output.md

@@ -0,0 +1,74 @@
+# Workflow and output contract
+
+Use this reference only when performing the full Customer Insights — Data or Customer Insights — Journeys review, or formatting the final answer.
+
+## Review domains
+
+Check these areas before giving a verdict:
+
+- Data source ingestion: source connectivity, table selection, schema mapping, refresh schedule
+- Data unification: source field mapping completeness, deduplication rule quality, match rule coverage, merge policy correctness
+- Unified profile quality: profile completeness, identity resolution evidence, known duplicates or splits
+- Segment design: rule logic, profile source validation, membership count reasonableness, refresh cadence, downstream usage
+- Measures and KPIs: definition correctness, data currency, calculation logic, usage in segments or journeys
+- Consent model: compliance profile configuration, purpose and topic hierarchy, contact point consent coverage, double opt-in status, legacy consent migration completeness
+- Journey design: trigger or segment entry configuration, branch logic, channel steps, personalization token validity, suppression and frequency cap configuration
+- Compliance posture: GDPR, CAN-SPAM, CASL, or other applicable regulation alignment; unsubscribe link presence; preference center availability
+- Stakeholder sign-off: named compliance owner and marketing operations lead approvals, dated
+- Post-launch monitoring: engagement metrics baseline, consent opt-out rate tracking, journey error rate monitoring
+
+## Safe workflow
+
+1. **Frame scope**
+   - Customer Insights workloads in scope (Data, Journeys, or both):
+   - Data sources being unified:
+   - Journey types in scope (segment-based, trigger-based, or both):
+   - Regulatory jurisdictions in scope (GDPR, CAN-SPAM, CASL, other):
+   - Explicit non-goals:
+
+2. **Collect evidence**
+   - Prefer documented artifacts: unification run logs, segment inspection exports, consent audit reports, journey test run results, compliance profile configuration, stakeholder sign-off records.
+   - Otherwise inspect sanitized user-provided summaries or official Dynamics 365 Customer Insights documentation.
+   - Label each finding as `live evidence`, `documented artifact`, `user-provided evidence`, `documentation-based`, or `inference`.
+
+3. **Stress-test risk**
+   - Have all data sources been ingested and schema-mapped before unification runs?
+   - Has the deduplication and match configuration been validated with sample profile results?
+   - Do segment rules reference correct profile attributes and are refresh schedules aligned with journey entry cadence?
+   - Has the consent model been configured with at least one compliance profile, purpose, and topic before journey publish?
+   - Is double opt-in configured on the relevant compliance profiles if required by applicable regulation?
+   - Has legacy consent been migrated from outbound marketing or external systems?
+   - Have all journey branches been tested including suppression, exit, and error paths?
+   - Has the compliance owner signed off on the consent model and unsubscribe path before production journey publish?
+   - What evidence is missing that would change the verdict?
+
+4. **Recommend the smallest safe action**
+   - Prefer additional unification runs over proceeding to production journeys with incomplete profiles, staged consent migration over bulk outreach, and journey test runs over direct production publish.
+   - If the safest action is to stop and complete consent model configuration or validate segment membership, say that plainly.
+   - Production journey publish and bulk outreach require live-guard escalation. Do not recommend production publish without explicit human approval from the marketing operations lead and compliance owner.
+
+## Output contract
+
+Return this structure:
+
+```markdown
+# D365 Customer Insights Review: <scope>
+## Executive verdict
+- Status: READY / READY WITH CONDITIONS / NOT READY / NEEDS EVIDENCE
+- Biggest risk:
+- Evidence level:
+## Scope and assumptions
+- Confirmed:
+- Unknown:
+- Out of scope:
+## Findings
+| Severity | Finding | Evidence | Why it matters | Minimum safe action |
+|---|---|---|---|---|
+## Recommended actions
+1. <action> — owner: <owner>, validation: <check>, rollback: <rollback>
+## Validation
+- Artifacts or checks to review:
+- Expected result:
+## Residual risk
+- <risk or explicit none>
+```

package/skills/microsoft/d365-customer-service-contact-center/SKILL.md

@@ -0,0 +1,62 @@
+---
+name: d365-customer-service-contact-center
+description: Review Dynamics 365 Customer Service and Contact Center operations across the case-to-resolution lifecycle — case management, unified routing, Omnichannel for Customer Service, queues, entitlements, service-level agreements (SLAs), knowledge management, and Copilot in Service. Use to improve case resolution time, routing accuracy, knowledge reuse, omnichannel consistency, and CSAT. Static review only; production routing-rule, SLA, and channel configuration changes are escalated.
+allowed-tools: Read Grep Glob
+metadata:
+  author: "github: Raishin"
+  version: "0.1.0"
+  updated: "2026-06-16"
+  category: operational
+---
+
+# D365 Customer Service & Contact Center
+
+## Purpose
+
+Act as the Dynamics 365 Customer Service reviewer who treats every misrouted case, breached SLA, stale knowledge base, and inconsistent channel experience as a CSAT and cost-to-serve risk until proven otherwise. Cover the case-to-resolution lifecycle from intake across channels through routing, resolution, and knowledge reuse.
+
+## When to use
+
+Use this skill for:
+
+- Case management: record creation/update rules, case resolution experience, status reasons, timeline
+- Unified routing: rules, skills-based routing, capacity profiles, work classification, queue assignment
+- Omnichannel for Customer Service: voice, chat, messaging channels, workstreams, agent experience
+- Queues, entitlements, and service scheduling
+- Service-level agreements (SLAs): KPIs, applicable-when conditions, success criteria, warning/failure actions, recalculation
+- Knowledge management: article authoring/curation/publishing, internal and external (SharePoint) search, AI suggestions
+- Copilot in Service: case/conversation summaries, ask-a-question, draft-a-response, knowledge integration
+- Customer Service KPIs: first-response/resolution time, SLA attainment, routing accuracy, deflection, CSAT
+
+Do not use this skill for:
+
+- Field Service work orders and scheduling (use d365-field-service-to-cash)
+- Sales pipeline/forecasting (use d365-sales-revenue-operations)
+- Customer Insights segmentation/journeys (separate skill)
+
+## Lean operating rules
+
+- Prefer current Microsoft Learn documentation for Customer Service, Omnichannel, unified routing, and SLA behavior. Administration is configured in the Copilot Service admin center (formerly Customer Service admin center); verify current naming.
+- Separate confirmed facts from inference. If SLA attainment or routing metrics were not provided, say so.
+- Challenge cases routed manually at scale, SLAs without warning actions, knowledge bases without curation/expiry, and channels configured without capacity profiles.
+- Keep answers scoped, reversible, and explicit about blockers or unknowns.
+- Load references only when needed.
+- Never ask for credentials, environment URLs, tenant IDs, connection strings, or customer data.
+
+## References
+
+Load these only when needed:
+
+- [Workflow and output contract](references/workflow-and-output.md) — use when executing the full Customer Service review or formatting the final answer.
+- [Safety checklist](references/safety-checklist.md) — use before any recommendation involving production routing rules, SLA configuration, or channel changes.
+- [Official sources](references/official-sources.md) — use when grounding Customer Service, Omnichannel, unified routing, or SLA behavior.
+
+## Response minimum
+
+Return, at minimum:
+
+- the scoped target and evidence level,
+- the main case-management, routing, SLA, knowledge, or omnichannel gaps,
+- the safest next actions,
+- validation or rollback notes where relevant,
+- the assumptions or blockers that prevent stronger conclusions.

package/skills/microsoft/d365-customer-service-contact-center/metadata.json

@@ -0,0 +1,30 @@
+{
+  "id": "d365-customer-service-contact-center",
+  "name": "D365 Customer Service & Contact Center",
+  "type": "skill",
+  "provider": "microsoft",
+  "harnesses": [
+    "codex",
+    "claude-code",
+    "cursor",
+    "gemini",
+    "kiro",
+    "other"
+  ],
+  "summary": "Review Dynamics 365 Customer Service and Contact Center across the case-to-resolution lifecycle: case management, unified routing, Omnichannel for Customer Service, queues, entitlements, SLAs, knowledge management, and Copilot in Service to improve resolution time, routing accuracy, knowledge reuse, and CSAT.",
+  "source_type": "original",
+  "official_docs": [
+    "https://learn.microsoft.com/dynamics365/customer-service/implement/overview",
+    "https://learn.microsoft.com/dynamics365/customer-service/administer/overview-cases",
+    "https://learn.microsoft.com/dynamics365/customer-service/administer/create-enhanced-sla",
+    "https://learn.microsoft.com/dynamics365/guidance/business-processes/case-to-resolution-overview",
+    "https://learn.microsoft.com/dynamics365/customer-service/administer/create-standard-sla"
+  ],
+  "security_notes": "Static review only. Do not modify production unified-routing rules, SLA configurations, channel/workstream setup, or knowledge publishing workflows; these are escalated to a qualified Customer Service administrator with a rollback plan. Do not ask for credentials, environment URLs, tenant IDs, connection strings, or customer data. Treat misrouted cases, SLAs without warning actions, stale knowledge bases, and uncapped channel capacity as CSAT and cost-to-serve risks until reviewed.",
+  "last_verified": "2026-06-16",
+  "path": "skills/microsoft/d365-customer-service-contact-center",
+  "author": "github: Raishin",
+  "version": "0.1.0",
+  "category": "operational",
+  "companion_agents": ["d365-customer-service-contact-center-agent"]
+}

package/skills/microsoft/d365-customer-service-contact-center/references/official-sources.md

@@ -0,0 +1,25 @@
+# Official sources
+
+Use this reference only when you need source grounding for Dynamics 365 Customer Service behavior or the case-to-resolution business process.
+
+## Microsoft Learn documentation
+
+Use these as starting points, not as proof of the user's live Customer Service configuration or operational metrics:
+
+- https://learn.microsoft.com/dynamics365/customer-service/implement/overview — Customer Service capabilities: cases, interactions, knowledge base, unified routing, omnichannel (incl. voice), SLAs, entitlements, dashboards. Administration via the Copilot Service admin center. Supports scope and KPI workflow steps.
+- https://learn.microsoft.com/dynamics365/customer-service/administer/overview-cases — Case management components: cases, activities, entitlements, knowledge articles, queues, Copilot case summaries, SLAs, record creation/update rules, routing rules. Supports the case-management review step.
+- https://learn.microsoft.com/dynamics365/customer-service/administer/create-enhanced-sla — Enhanced SLAs: SLA KPIs, applicable-when conditions, success criteria, warning/failure actions. Supports the SLA review step.
+- https://learn.microsoft.com/dynamics365/customer-service/administer/create-standard-sla — Standard SLAs and the available SLA workflow actions. Supports the SLA design and pushback steps.
+- https://learn.microsoft.com/dynamics365/guidance/business-processes/case-to-resolution-overview — Case to resolution end-to-end process and featured capabilities: Omnichannel, unified routing, knowledge base, Microsoft Teams collaboration, AI suggestions, Customer Insights, Customer Voice. Supports the end-to-end lifecycle workflow.
+
+## Certification alignment (verify currency)
+
+- Exam MB-230: Microsoft Dynamics 365 Customer Service Functional Consultant — skills measured updated as of March 11, 2026 (verify status before citing): case management (51–55%), representative experience and routing (25–30%), extend Customer Service (15–20%).
+
+## Terminology note
+
+The Customer Service admin center was renamed the Copilot Service admin center. Some Learn articles may use either name. Verify current naming before presenting configuration locations.
+
+## Grounding rule
+
+Official documentation explains Customer Service behavior. It does not prove the user's actual SLA attainment, routing accuracy, knowledge freshness, channel configuration, or CSAT. Prefer exported reports, sanitized user-provided metrics, or read-only evidence for current-state claims.

package/skills/microsoft/d365-customer-service-contact-center/references/safety-checklist.md

@@ -0,0 +1,34 @@
+# Safety checklist
+
+Use this reference before any recommendation involving production unified-routing rules, SLA configuration, channel/workstream setup, or knowledge publishing in Dynamics 365 Customer Service.
+
+## Non-negotiables
+
+- Never ask users to paste credentials, tenant IDs, environment URLs, connection strings, certificates, or customer data into chat.
+- Use exported reports or sanitized user-provided evidence for current-state claims; otherwise use documentation and label the evidence level.
+- Do not invent SLA attainment percentages, case volumes, routing distributions, or CSAT scores.
+- Require explicit human approval before recommending any production routing-rule, SLA, channel, or knowledge-publishing change.
+- Use current official Microsoft Learn documentation for Customer Service, Omnichannel, and unified routing behavior.
+- Keep recommendations least-change, reversible, and scoped to the domain in question.
+
+## Stress checks
+
+- What cases are misrouted, reassigned repeatedly, or manually triaged at scale?
+- Which SLAs lack warning actions, realistic business calendars, or recalculation-on-reopen settings?
+- Is the knowledge base curated, current, and de-duplicated, or stale?
+- Are omnichannel capacity profiles set to prevent agent overload?
+- What rollback exists if a routing-rule or SLA change misbehaves in production?
+
+## Evidence labels
+
+Use `documented artifact`, `user-provided evidence`, `documentation-based`, or `inference`. Documentation alone never proves the user's actual SLA attainment, routing accuracy, knowledge freshness, or CSAT.
+
+## Live-guard gate
+
+The following actions require explicit human confirmation and are out of scope for automated execution:
+
+- Modifying production unified-routing rules, work classification, or capacity profiles
+- Changing production SLA KPIs, success criteria, or warning/failure actions
+- Configuring or reconfiguring production channels/workstreams (voice, chat, messaging)
+- Publishing, unpublishing, or bulk-editing production knowledge articles
+- Executing bulk case reassignment, status changes, or queue moves

package/skills/microsoft/d365-customer-service-contact-center/references/workflow-and-output.md

@@ -0,0 +1,66 @@
+# Workflow and output contract
+
+Use this reference only when performing the full Dynamics 365 Customer Service & Contact Center review or formatting the final answer.
+
+## Review domains
+
+Check these areas before giving a verdict:
+
+- Case management: record creation/update rules, resolution experience, status reasons, timeline configuration, security roles
+- Unified routing: rule-based vs. skills-based routing, work classification, capacity profiles, queue/agent assignment
+- Omnichannel: configured channels (voice, chat, messaging), workstreams, agent experience, capacity
+- Queues and entitlements: queue design, entitlement terms, service scheduling
+- SLAs: KPIs, applicable-when conditions, success criteria, warning/failure actions, recalculation on terminal status
+- Knowledge management: authoring/curation/publishing, internal and external (SharePoint) search, AI suggestions, article freshness/expiry
+- Copilot in Service: case/conversation summaries, ask-a-question, draft-a-response, knowledge sources
+- KPIs: first-response/resolution time, SLA attainment, routing accuracy, self-service deflection, CSAT
+
+## Safe workflow
+
+1. **Frame scope**
+   - Area in scope (case management / routing / omnichannel / SLA / knowledge):
+   - Licensing and channels in use (Customer Service, Contact Center, Omnichannel voice):
+   - Required outcome (resolution time / routing accuracy / SLA attainment / knowledge reuse / CSAT):
+   - Explicit non-goals:
+
+2. **Collect evidence**
+   - Prefer exported reports: case aging, SLA attainment, routing distribution, knowledge usage, CSAT survey results.
+   - Otherwise inspect sanitized user-provided summaries or official Microsoft Learn documentation.
+   - Label each finding as `documented artifact`, `user-provided evidence`, `documentation-based`, or `inference`.
+
+3. **Stress-test risk**
+   - What cases are misrouted or manually triaged at scale?
+   - Which SLAs lack warning actions or realistic business-hours calendars?
+   - Is the knowledge base curated and current, or stale and duplicative?
+   - Are channels configured with capacity profiles to prevent agent overload?
+   - What evidence is missing that would change the verdict?
+
+4. **Recommend the smallest safe action**
+   - Prefer configuration and process fixes over channel/routing-engine changes.
+   - Production routing-rule, SLA, and channel configuration changes require live-guard escalation with a rollback plan.
+
+## Output contract
+
+Return this structure:
+
+```markdown
+# D365 Customer Service & Contact Center Review: <scope>
+## Executive verdict
+- Status: HEALTHY / HEALTHY WITH RISKS / AT RISK / NEEDS EVIDENCE
+- Biggest risk:
+- Evidence level:
+## Scope and assumptions
+- Confirmed:
+- Unknown:
+- Out of scope:
+## Findings
+| Severity | Finding | Evidence | Why it matters | Minimum safe action |
+|---|---|---|---|---|
+## Recommended actions
+1. <action> — owner: <owner>, validation: <check>, rollback: <rollback>
+## Validation
+- Reports or checks to review:
+- Expected result:
+## Residual risk
+- <risk or explicit none>
+```

package/skills/microsoft/d365-data-migration-cutover/SKILL.md

@@ -0,0 +1,58 @@
+---
+name: d365-data-migration-cutover
+description: Review Dynamics 365 data migration planning and go-live cutover readiness. Enforces mock migration evidence, data quality gates, staging table validation, reconciliation controls, cutover runbook completeness, rollback plan, and owner sign-off before production migration. Refuses to bless production cutover without reconciliation evidence and rollback plan. Production data migration is live-guard gated and requires escalation.
+allowed-tools: Read Grep Glob
+metadata:
+  author: "github: Raishin"
+  version: "0.1.0"
+  updated: "2026-06-16"
+  category: data
+---
+
+# D365 Data Migration & Cutover
+
+## Purpose
+
+Act as the Dynamics 365 data migration and cutover reviewer who treats every unvalidated legacy data extract, failed or skipped mock migration, missing reconciliation evidence, and untested rollback plan as a production go-live blocker until proven otherwise.
+
+## When to use
+
+Use this skill for:
+
+- Data migration strategy review (entity selection, migration sequence, full vs. delta load approach)
+- Data Management Framework usage: data entities, data packages, import/export jobs, staging table validation
+- Legacy data quality assessment and cleansing planning
+- Mock migration (dry run) planning, execution, and results review
+- Reconciliation control design: record counts, field-level sampling, business user validation
+- Cutover runbook review: sequenced tasks, owners, durations, dependencies, go/no-go criteria
+- Rollback plan assessment: rollback triggers, rollback owner, rollback time window, rollback validation
+- Production data migration authorization and go-live cutover readiness sign-off
+- Post-migration hyper-care: data validation, issue triage, escalation path
+
+## Lean operating rules
+
+- Prefer current Microsoft Learn documentation for Dynamics 365 data management framework behavior, data entities, and cutover strategy. Use the per-skill facts and sources in `references/official-sources.md` for grounding.
+- Separate confirmed facts from inference. If a mock migration was not completed or reconciliation evidence was not provided, say so explicitly.
+- Challenge missing mock migration evidence, unvalidated legacy data, record-count-only reconciliation, missing rollback plans, and cutover authorizations without business owner sign-off.
+- Keep answers scoped, reversible, and explicit about blockers or unknowns.
+- Load references only when needed; do not pull all deep guidance into short answers.
+- Never ask for credentials, connection strings, environment URLs, tenant IDs, or customer data.
+
+## References
+
+Load these only when needed:
+
+- [Workflow and output contract](references/workflow-and-output.md) — use when executing the full migration or cutover review, or formatting the final answer.
+- [Safety checklist](references/safety-checklist.md) — use before any recommendation involving production data migration, cutover authorization, or rollback execution.
+- [Official sources](references/official-sources.md) — use when grounding Data Management Framework behavior, data entity usage, or cutover strategy guidance.
+- [Data Migration & Cutover Guide](references/data-migration-cutover-guide.md) — use for domain-specific failure modes, safe review workflow, verification targets, and pushback criteria.
+
+## Response minimum
+
+Return, at minimum:
+
+- the scoped migration target and evidence level,
+- the main data quality issues, mock migration gaps, reconciliation risks, or cutover runbook gaps,
+- the safest next actions,
+- validation or rollback notes where relevant,
+- the assumptions or blockers that prevent stronger conclusions.

package/skills/microsoft/d365-data-migration-cutover/metadata.json

@@ -0,0 +1,32 @@
+{
+  "id": "d365-data-migration-cutover",
+  "name": "D365 Data Migration & Cutover",
+  "type": "skill",
+  "provider": "microsoft",
+  "harnesses": [
+    "codex",
+    "claude-code",
+    "cursor",
+    "gemini",
+    "kiro",
+    "other"
+  ],
+  "summary": "Review Dynamics 365 data migration planning and go-live cutover readiness. Enforces mock migration evidence, data quality gates, staging table validation, reconciliation controls, cutover runbook completeness, rollback plan, and owner sign-off before production migration. Detects dirty legacy data, failed mock migrations, weak reconciliation, missing rollback paths, and go-live chaos risks. Refuses to bless production cutover without reconciliation evidence and rollback plan. Production data migration is live-guard gated and requires escalation.",
+  "source_type": "original",
+  "official_docs": [
+    "https://learn.microsoft.com/dynamics365/guidance/implementation-guide/prepare-to-go-live",
+    "https://learn.microsoft.com/dynamics365/guidance/implementation-guide/prepare-go-live-checklist",
+    "https://learn.microsoft.com/dynamics365/guidance/implementation-guide/prepare-go-live-cutover-strategy",
+    "https://learn.microsoft.com/dynamics365/fin-ops-core/dev-itpro/data-entities/data-entities-data-packages",
+    "https://learn.microsoft.com/dynamics365/fin-ops-core/fin-ops/data-entities/data-import-export-job",
+    "https://learn.microsoft.com/training/modules/prepare-data-migration-finance-operations/",
+    "https://learn.microsoft.com/power-platform/architecture/key-concepts/data-migration/cut-over-planning"
+  ],
+  "security_notes": "Never approve production data migration or cutover without documented evidence of at least one completed mock migration, reconciliation sign-off comparing source and target record counts, and a tested rollback plan. Production data migration is live-guard gated and must be escalated to a human implementation lead and business data owner before execution. Data quality issues discovered during staging table validation must be resolved and re-validated before migration to production. Do not accept record count matching alone as reconciliation evidence; require field-level sampling and business user validation. Do not ask for credentials, connection strings, environment URLs, tenant IDs, or customer data. Treat every unvalidated legacy data extract, failed mock migration, and missing rollback plan as a production go-live blocker.",
+  "last_verified": "2026-06-16",
+  "path": "skills/microsoft/d365-data-migration-cutover",
+  "author": "github: Raishin",
+  "version": "0.1.0",
+  "category": "data",
+  "companion_agents": ["d365-data-migration-cutover-agent"]
+}

package/skills/microsoft/d365-data-migration-cutover/references/data-migration-cutover-guide.md

@@ -0,0 +1,76 @@
+# Data Migration & Cutover Guide
+
+Use this reference for Dynamics 365 data migration failure modes, safe review workflow, Data Management Framework patterns, reconciliation design, cutover runbook requirements, verification targets, and pushback criteria.
+
+## What people get wrong
+
+The lazy story is:
+
+> Run an export from the legacy system, import it into Dynamics 365 using the Data Management workspace, check the record count, and you are done.
+
+Wrong. Record count matching is a necessary but not sufficient reconciliation control. Field-level data quality issues, staging table transformation errors, dependency sequencing failures, and financial balance discrepancies are invisible to record count checks alone. A single failed mock migration that is declared "good enough" is the leading predictor of production cutover chaos.
+
+Common bad assumptions:
+
+- One mock migration is sufficient regardless of error rate or elapsed time.
+- Staging table errors that were manually resolved in the test environment do not need to be re-validated after fix.
+- Record count reconciliation is sufficient to prove data integrity.
+- The cutover window is generous enough to absorb a 2x slowdown from production environment differences.
+- Legacy data cleansing can be completed during the cutover window.
+- Rollback is simple and can be planned ad hoc if needed.
+- Delta loads between the final mock migration and production cutover do not need a separate test run.
+
+## Data migration failure modes
+
+- Legacy data quality issues (duplicates, nulls in required fields, invalid codes, format mismatches) are discovered during production migration after the cutover window has started.
+- Staging table errors not resolved before production promotion cause partial entity loads, creating referential integrity violations in target tables.
+- Entity dependency sequence not respected: transactional data imported before master data causes failed lookups.
+- Migration throughput measured in development or sandbox environment does not reflect production environment capacity, causing the cutover window to be exceeded.
+- Delta load strategy not designed: data changed in the legacy system between the final mock migration and production cutover is not captured, leaving gaps.
+- Rollback plan not tested: when rollback is needed, the steps are unclear, the owner is unknown, and the rollback itself fails.
+- Reconciliation performed only by the implementation team, not validated by business data owners who know the expected values.
+
+## High-risk data domains (examples)
+
+- **Chart of accounts and financial dimensions**: must be migrated and validated before any financial transaction data; errors here cascade across all financial reporting.
+- **Customer and vendor master data**: deduplication failures create billing, payment, and compliance issues post-go-live.
+- **Open purchase orders and open invoices**: in-flight transactions must be captured completely or business operations halt at go-live.
+- **Fixed asset master and net book values**: incorrect opening balances trigger depreciation calculation errors.
+- **Inventory on-hand balances**: quantity or cost inaccuracies surface immediately at go-live in procurement and production.
+- **Bank account and payment method configuration**: missing or incorrect data blocks payment processing from day one.
+
+## Minimum safe migration review workflow
+
+1. Confirm scope: Dynamics 365 workloads, legal entities, data domains, legacy sources, go-live date, cutover window.
+2. Review entity selection and dependency sequence: all required entities identified, sequenced in dependency order.
+3. Confirm mock migration evidence: at least one completed mock migration with documented staging error counts, elapsed time per entity, and reconciliation totals.
+4. Review data quality plan: legacy data cleansing completed or in progress, known issues logged with mitigations.
+5. Review reconciliation design: record counts, field-level sampling plan, financial balance checks, business user validation sign-off.
+6. Review cutover runbook: task sequence, owners, durations, system freeze point, delta load steps, go/no-go checkpoint.
+7. Review rollback plan: trigger criteria, rollback owner, rollback steps, rollback validation, rollback time estimate.
+8. Confirm stakeholder sign-off: business data owner and implementation lead approvals documented.
+9. Provide minimum-safe-action recommendation scoped to highest-severity migration gaps.
+10. Require live-guard escalation for any production migration authorization.
+
+## Verification targets
+
+- Mock migration results: completed runs, staging error count, elapsed time, entity coverage
+- Data quality: cleansing status, known issues log, deduplication results
+- Reconciliation: record counts, field samples, financial balances, business user validation
+- Cutover runbook: task sequence, owners, durations, go/no-go criteria, rollback section
+- Rollback plan: trigger criteria, owner, steps, validation, time estimate
+- Delta load strategy: designed, tested, sequenced relative to freeze point
+- Stakeholder sign-off: business data owner and implementation lead named and dated
+
+## When to push back
+
+Push back if the user asks to:
+
+- approve production migration without at least one completed mock migration with documented results
+- accept record count matching alone as sufficient reconciliation evidence
+- proceed to production cutover with unresolved staging table errors
+- approve a cutover window that has no throughput buffer based on measured migration performance
+- proceed without a tested rollback plan that includes a named rollback owner and trigger criteria
+- sign off on reconciliation without business data owner validation
+- skip the go/no-go checkpoint before starting production migration
+- authorize post-migration legacy data deletion before business users have validated the target environment