Start from a guided prompt
Use Quick Starts or search 94 curated prompts, fill parameters, copy the rendered request, and follow evidence-backed suggested next prompts.
Discoverability & Guided Investigation
Start from DV Quick Run's 94-prompt guided library or ask directly in GitHub Copilot. The deterministic, extension-owned Local MCP surface can understand architecture, return the live Operational Profile + DVQR Score, run bounded queries, use Custom API Intelligence, or—on Pro—continue persisted investigations through evidence, readiness, and bounded Mini RCA.

What it delivers
Use Quick Starts or search 94 curated prompts, fill parameters, copy the rendered request, and follow evidence-backed suggested next prompts.
Retrieve the canonical live profile and calibrated DVQR Score through Free MCP with explicit evidence, arithmetic, freshness, and interpretation boundaries.
Prepare a real record investigation with persisted subject, explicit target, deterministic strategy, and explicit intent confirmation.
Show confidence, evidence provenance, reasons, runtime status, and remaining uncertainty.
Separate business anchors, supporting coordination records, and downstream work items.
Resolve navigation properties and query shapes from active-environment metadata.
Execute target-aware relationship candidates hop-by-hop, record observed viability, and preserve the exact point where continuation becomes empty.
Save an explicitly verified route as workspace knowledge, surface it first later, and reuse the exact saved relationships from a real source record.
Reconcile the current evidence fingerprint, expose stale or missing checkpoints, and generate a bounded evidence-backed Mini RCA on Pro.
Prompt Library · guided discovery
Quick Starts help newcomers begin from common outcomes, while categories, search, Free/Pro filters, parameterised rendering, and suggested next prompts support deeper users. The Prompt Library is a catalogue and guidance layer—not a second reasoning engine.

New in v0.16.0
When DVQR finds and runtime-validates a useful route, it can offer to save that exact relationship path. Preferred paths remain evidence-bounded: they surface first later, but metadata-derived alternatives stay visible.
Find metadata-valid candidates and prove the asserted route against a bounded real record.
Persist only after user confirmation as inspectable JSON under .dvforgelab/dvqr/business-paths.
Surface saved business knowledge first while keeping other valid routes available.
Start from a real source record and continue through the exact saved relationships; empty frontiers stop safely.
Workspace-owned lifecycle
Select an environment, enable the Local MCP server for the workspace, sign in with Azure CLI, and let VS Code start the stdio process on demand. The Hub keeps environment, mode, authentication, lifecycle, and the Free/Pro MCP boundary visible at a glance.

Free vs Pro MCP
The tool split is deliberately visible in the DV Quick Run Hub so users can see the real MCP capability difference rather than guessing from plan names.
Free tools
30
Foundational understanding, bounded queries, metadata intelligence and basic eligible Custom API workflows.
Pro-only tools
32
Managed investigation continuity, runtime evidence, readiness, Mini RCA and deeper investigation acceleration.
Total with Pro
62
Every Free tool plus the complete current Pro MCP surface.
Available
Not exposed
Prompt playground
Try foundational understanding or move directly into a persisted Pro investigation. The prompt describes the outcome; DVQR keeps the evidence and execution boundaries explicit.