DV Quick Run

Understand Dataverse applications. Investigate operational behaviour. Stay grounded in evidence.

DV Quick Run is an extension-owned Local MCP server and metadata-aware Dataverse investigation workbench for VS Code. Ask Dataverse questions in natural language, discover business architecture, run bounded queries, follow real relationships, preserve verified business paths, and continue into evidence-backed investigation without leaving the editor.

Read-only MCP foundations · Deterministic metadata · Evidence-backed conclusions · Human authority

Website · Pricing · GitHub Discussions

What's New in v0.16.1

MCP Security Hardening I + Guided Traversal Business Path Capture

v0.16.1 hardens the model ↔ MCP ↔ DVQR boundary and completes the path from bounded traversal evidence into reusable Managed Business Paths.

validate → execute bounded → verify exact route → save/reverify → reuse safely

• Server-side MCP authority — registered schemas, capability allow-lists, environment binding, bounds and entitlement remain application-owned rather than prompt-controlled.
• Untrusted-input hardening — malformed identifiers, unsafe environment URLs, fabricated control fields, path/file manipulation and credential-shaped output are rejected or redacted at the DVQR boundary.
• Explicit Save / Verify — a useful Guided Traversal route can be deliberately captured; identical canonical routes retain the same Business Path ID and refresh verification evidence instead of creating duplicates.
• Evidence-aware reuse — saved guidance is revalidated against current metadata before reuse; Saved, runtime-verified and Preferred remain distinct states.
• Empty-frontier scope guard — a saved-path run that reaches no continuation terminates its investigation scope server-side. Automatic OData, alternate-route, target-expansion or probe broadening cannot silently continue it.
• Auditable new-scope transition — broader follow-up investigation requires an explicit Business Path scope transition before DVQR will continue.
• Accurate null-navigation evidence — null/204 singleton navigation results are treated as zero landed records rather than apparent returned data.

Managed Business Paths remain workspace guidance, not Dataverse truth: current metadata, bounded runtime evidence, user preference and causality stay separate.

Discoverability remains built in

The 94-prompt Prompt Library (69 Free + 25 Pro) remains the guided entry point to DVQR's Local MCP capabilities, including Operational Profile + calibrated DVQR Score, relationship intelligence, bounded traversal and managed investigation.

Start in Three Ways

Open the two main orientation surfaces from the Command Palette:

DV Quick Run: Open Prompt Library
DV Quick Run: Open Hub

What DV Quick Run Can Do

1. Understand Dataverse

Translate structural metadata and bounded runtime evidence into an explainable view of an unfamiliar application.

• discover business capabilities and operational anchors
• identify Core Domain, Coordination, Execution, Governance and Platform layers
• explain why entities and routes were ranked
• distinguish metadata recommendation from runtime-observed workflow
• preserve confidence, provenance and uncertainty

2. Query & Explain

Work directly with Dataverse query surfaces in VS Code.

• natural-language OData through Local MCP
• OData and FetchXML execution
• $batch workflows
• Query-by-Canvas refinement
• Query Doctor and metadata-aware suggestions
• Query Explain
• preview-first Smart PATCH in the editor

3. Navigate Real Business Relationships

Move from metadata-valid relationships to data-viable traversal.

• relationship discovery and ranking
• Guided Traversal using returned rows
• Relationship Graph
• runtime path verification
• Managed Business Paths
• Preferred-path reuse with exact saved hops
• $batch traversal replay

4. Investigate Operational Behaviour

Use bounded evidence instead of speculative diagnosis.

• Result Viewer
• Execution Insights
• Operational Profiles
• DVQR Score
• Access Context for users, teams, roles, business units and application identities
• persisted Pro managed investigations
• evidence acquisition, readiness and Mini RCA

5. Compare & Reconstruct

Preserve evidence across environments and time.

• Evidence Workspace and Snapshot Library
• Cross-Environment Diff
• Timeline Reconstruction
• Timeline Graph
• Audit Evidence Enrichment
• findings, verification and handoff workflows
• HTML/PDF investigation reports

6. Discover & Govern Dataverse Capabilities

Inspect supported Custom APIs and Dataverse operations with explicit execution boundaries.

• Capability Explorer
• bound and unbound operation discovery
• metadata-backed request shaping
• preview-first supported execution
• explicit confirmation
• access-aware discovery
• execution diagnostics and investigation continuation

See It in Action

Ask Dataverse in natural language

Find related tables deterministically

Explain OData in plain English

Explore query results

Understand an entity's operational footprint

Free and Pro

DV Quick Run follows an open-core model. Foundational Dataverse understanding remains accessible; Pro adds advanced investigation acceleration and persistence.

A 14-day Pro Trial is available for teams that want to evaluate the full investigation workflow.

5-Minute Quick Start

A. Run DV Quick Run normally

1. Install DV Quick Run in VS Code.

2. Configure and select a Dataverse environment.

3. Run a query such as:

   contacts?$top=10
   

4. Open DV Quick Run: Open Hub whenever you need orientation.

B. Enable Local MCP

1. Select the Dataverse environment to expose to the local server.

2. Run:

   DV Quick Run: Enable Local MCP Server
   

3. Sign into the environment tenant with Azure CLI. For a tenant without an Azure subscription:

   az login --tenant <tenant-id> --allow-no-subscriptions
   

4. Open GitHub Copilot Chat and make sure the DV Quick Run MCP tools are enabled.

5. Open DV Quick Run: Open Prompt Library, or ask directly:

   Using DV Quick Run, show me what I can investigate in this Dataverse environment and recommend where to start.
   

DV Quick Run remembers MCP enablement per workspace. VS Code starts the extension-owned local stdio server on demand.

Local MCP

DV Quick Run owns its local MCP lifecycle inside the extension. The Hub shows the selected environment, mode, tool count, authentication guidance and traffic-light health state.

The foundational MCP surface is deliberately bounded. It supports metadata understanding, relationship intelligence, natural-language querying and evidence acquisition without turning the assistant into an unrestricted Dataverse administrator.

Typical prompts:

Using DV Quick Run, find tables related to customers.

Using DV Quick Run, explain this OData query:
accounts?$select=name,revenue&$filter=statecode eq 0&$orderby=name asc&$top=10

Using DV Quick Run, starting only from Contact metadata, discover the business capabilities and explain where operational work is coordinated and performed.

Using DV Quick Run, find and runtime-verify a relationship path from account to task.

Business Architecture Understanding

Operational Workflow Intelligence separates structural evidence into:

• Core Domain — principal business, service, case, plan or request concepts
• Coordination — journey, process, routing and orchestration records
• Execution — tasks, activities and downstream work items
• Governance — eligibility, consent, approval, safety and control records
• Platform — plugins, flows, asynchronous jobs and integration participation

Strong structural evidence does not prove that a particular record participated. Runtime evidence remains investigation-scoped and does not overwrite metadata confidence.

Managed Business Paths

Managed Business Paths turn verified traversal knowledge into a reusable workspace asset.

metadata-valid path
        ↓
runtime verification
        ↓
explicit save
        ↓
Preferred workspace path
        ↓
record-scoped exact-hop Guided Traversal

A Preferred path is top visible, not exclusive. Metadata-derived alternatives remain available. Preference records useful team knowledge; it does not replace Dataverse metadata truth or prove universal business authority.

Saved paths live under:

.dvforgelab/dvqr/business-paths

The runtime frontier is intentionally bounded:

landed record IDs → exact next relationship → landed record IDs

If a hop returns no rows, traversal stops. DVQR does not convert an empty frontier into a table-wide query.

Result Viewer & Guided Traversal

The Result Viewer is the main interactive surface for exploring query results.

start simple → run → explore → refine → investigate → verify

Use it to:

• view records as table or JSON
• search and inspect results
• refine queries
• launch record investigation
• open relationship navigation
• continue Guided Traversal
• inspect operational context
• preview compatible bound Actions
• export evidence and supported handoff artifacts

Guided Traversal carries actual landed records from hop to hop. Managed Preferred paths can reuse the exact saved route from a supplied source record.

Operational Profiles & DVQR Score

Operational Profiles describe the operational footprint of a Dataverse entity before deeper troubleshooting.

Profiles can surface:

• plugin orchestration density
• relationship complexity
• metadata footprint
• async participation
• Power Automate involvement
• workflow participation
• managed-state context

Profiles are entity-scoped, user-triggered, evidence-backed, bounded and advisory-only. The DVQR Score is a calibrated investigation aid, not a risk or root-cause score.

Access Context

Access Context investigates bounded identity participation for:

• users and application identities
• teams
• roles
• business units

It can show business-unit context, direct and inherited participation, team membership and supporting evidence.

It does not simulate RBAC, calculate effective record access, generate privilege matrices or infer security risk.

Managed Investigation

Pro managed investigations preserve professional investigation continuity rather than treating each prompt as an isolated answer.

prepare & confirm
      ↓
acquire bounded evidence
      ↓
assess readiness
      ↓
generate bounded Mini RCA
      ↓
verify / hand off

The evidence model keeps supported, weakened and unresolved hypotheses distinct and preserves gaps rather than filling them with speculation.

Evidence Workspace, Comparison & Timeline

DV Quick Run uses a local DV ForgeLab Evidence Workspace for investigation artifacts and reconstruction handoffs.

.dvforgelab
└─ dvqr
   ├─ business-paths
   ├─ comparisons
   ├─ reports
   └─ snapshots

Snapshot Library

Snapshot Library coordinates saved operational snapshots and comparison/timeline selection.

Cross-Environment Diff

Compare compatible snapshots across environments and review grouped operational drift with evidence references, verification state and handoff context.

Timeline Reconstruction

Select 3+ compatible snapshots from the same environment and entity to reconstruct snapshot-bounded intervals and first-observed drift.

DVQR can surface:

• Timeline Graph
• provider and significance distributions
• Timeline Trust
• Timeline Findings Summary
• Timeline Investigation Handoff
• optional Audit Evidence Enrichment

Timeline findings describe when drift was first observed between captures. They do not claim an exact change time, root cause, human responsibility or remediation status.

Capability Explorer & Governed Execution

Capability Explorer discovers and explains supported Dataverse operations and Custom APIs.

It distinguishes:

• Functions vs Actions
• bound vs unbound operations
• public vs private visibility
• preview-ready vs inspect-only capability
• execution eligibility
• operational-impact cautions

Supported execution follows:

preview → explicit confirmation → execution → inspect result → investigate evidence

Custom API metadata is discovery truth; OData metadata is execution-exposure truth; bound route metadata is execution-route truth; the active environment remains the execution authority boundary.

AI-related operation execution is blocked by default unless explicitly allowed by policy, and generated output still requires human review.

DV ForgeLab Ecosystem

DV Quick Run investigates. Other DV ForgeLab utilities reconstruct or execute focused changes through explicit handoff artifacts.

Supported ecosystem handoffs include:

• DVBUR — focused bulk upsert artifacts
• DVAF — supported attribute reconstruction intent
• DVIM — identity management artifacts
• DVCE — choice artifacts
• DVEVM — environment variable artifacts

Investigation and reconstruction remain separate concerns.

Guardrails

DV Quick Run favours explicit, preview-first, user-controlled workflows.

Key boundaries include:

• bounded queries and explicit execution context
• no silent Managed Business Path persistence
• no broadening after an empty traversal frontier
• no causal claims from participation alone
• no effective-access claims from Access Context participation
• no cross-environment timeline reconstruction
• no treatment of first-observed drift as exact historical time
• no treatment of audit rows as proof of causality
• no treatment of reports or exports as approval/certification
• no treatment of reconstruction artifacts as automatic remediation
• explicit confirmation for supported execution-capable workflows

DV Quick Run does not treat generated or AI-assisted responses as operational truth. Human verification remains the authority boundary.

Pro Activation

DV Quick Run is available in Free and Pro editions.

Pricing: https://www.dvquickrun.com/pricing

Purchase: https://dvforgelab.lemonsqueezy.com

Activate Online Pro from the Command Palette:

DV Quick Run: Activate Pro License

Inspect entitlement:

DV Quick Run: License Status

Restricted or disconnected environments can use signed Offline Pro licensing:

DV Quick Run: Import Offline License

Who Is This For?

• Dataverse / Dynamics 365 developers
• Power Platform engineers
• integration and API developers
• support engineers investigating Dataverse behaviour
• consultants working across complex Dataverse environments
• teams onboarding into unfamiliar Dataverse solutions

Why DV Quick Run?

Because useful Dataverse investigation is more than running a query:

understand → query → verify → investigate → compare → explain → hand off

DV Quick Run reduces tool switching while keeping metadata truth, runtime evidence, user preference, execution authority and human judgement explicitly separated.

Community & Feedback

GitHub Discussions: https://github.com/yongjinsim-sudo/dv-quick-run/discussions

Official website: https://www.dvquickrun.com

Use Discussions to report bugs, suggest features, share workflow feedback, discuss investigation patterns and submit evidence edge cases.

Development

npm install
npm run compile

Press F5 to run the extension.

For direct developer MCP verification after compiling:

npm run mcp:start

For interactive inspection:

npm run mcp:inspect

Azure CLI authentication must already be available through az login.

License & Open-Core Model

DV Quick Run follows an open-core model.

The MIT-licensed core preserves foundational Dataverse understanding workflows. Proprietary Pro modules provide advanced investigation acceleration, persistence, comparison, timeline, reporting and governed execution capabilities.

/src/core
MIT open-core functionality

/src/pro
Private proprietary acceleration modules
Not included in the public repository or MIT grant

Foundational operational understanding remains accessible. Commercial acceleration funds continued development.