Verity Framework — Delivery Item Prompt Generator

Generate structured AI development prompts directly from your Azure DevOps work items.

The Verity Framework ADO Extension reads the five Delivery Item fields defined by the IdeaRoost Verity Framework and generates a structured, ready-to-use prompt for Claude Code, Cursor, GitHub Copilot Workspace, or any AI development tool.

The Problem It Solves

When engineers use AI development tools, they typically start from a blank context — a vague task description or nothing at all. The AI has no awareness of the problem being solved, the scope constraints, the production risk level, or the specific hardening evidence required before the feature can ship.

The result is AI-generated code that may work functionally but lacks the context needed to make it production-ready.

How It Works

The extension adds a Verity Prompt tab to your Azure DevOps work items. When your team populates the five Verity Framework fields during planning, the extension generates a structured prompt that carries that context into the AI development tool:

Field	What It Captures
VF Intent	The problem and who experiences it
VF Value	The outcome if the problem is solved
VF Appetite	The time/resource ceiling (not an estimate)
VF Trust Criteria	Gate 2 hardening requirements before shipping
VF Failure Scope	Narrow / Moderate / Broad / Systemic risk level

The generated prompt includes:

Full problem context from Intent and Value
Scope constraints from Appetite
Specific implementation requirements from Trust Criteria
Risk-level constraints based on Failure Scope
A Definition of Done checklist mapped from Trust Criteria
Prompt injection protection on all user-supplied field content

Field Validation

The extension validates all five fields before generating a prompt and provides specific guidance when fields are incomplete or incorrectly formatted:

Intent flags solution-description language (Build, Create, Add...)
Appetite rejects priority words (high, medium) and requires time expressions (3 days, 0.5 sprint)
Trust Criteria flags vague language that cannot be verified at Gate 2

A field quality indicator shows the status of each field at a glance. The prompt is gated until all fields meet the quality threshold.

Setup (20 Minutes, Project Admin Access)

The extension works with five custom fields added to your existing User Story or Product Backlog Item work item type. No new work item type required.

Step 1: Create an inherited process from your current process (Organization Settings → Boards → Process)

Step 2: Add five custom fields to your User Story:

Field Name	Reference Name	Type
VF Intent	Custom.VFIntent	Multi-line text
VF Value	Custom.VFValue	Multi-line text
VF Appetite	Custom.VFAppetite	Single-line text
VF Trust Criteria	Custom.VFTrustCriteria	Multi-line text
VF Failure Scope	Custom.VFFailureScope	Picklist: Narrow, Moderate, Broad, Systemic

Step 3: Switch your project to the inherited process

Step 4: Install this extension — the Verity Prompt tab appears on all User Story work items automatically

Full step-by-step setup instructions, including field format guidance with good/bad examples, are built into the extension itself — open the Setup Instructions panel on the Verity Prompt tab after installing.

Configurable Field Names

If your team uses different field reference names, configure them in the Verity Framework Settings hub (Organization Settings → Extensions → Verity Framework Settings). This hub requires Collection/Organization Administrator access, since the field-name mapping applies across every project in the organization.

The Verity Framework

This extension is part of the IdeaRoost Verity Framework — a ground-up delivery methodology for AI-augmented software teams. The framework defines five principles, four roles, four ceremonies, three production gates, and six metrics designed for teams where AI is doing most of the building.

Learn more about the framework →

Support & Legal

Questions, issues, or feedback: Hello@idearoost.com

Built by IdeaRoost — AI accelerates execution. Experience determines quality.