38-ado-poc ? EVA ADO Command Center

EVA Ecosystem Integration

Tool	Purpose	How to Use
37-data-model	Single source of truth for all project entities	GET http://localhost:8010/model/projects/38-ado-poc
29-foundry	Agentic capabilities (search, RAG, eval, observability)	C:\eva-foundry\eva-foundation\29-foundry
48-eva-veritas	Trust score and coverage audit	MCP tool: audit_repo / get_trust_score
07-foundation-layer	Copilot instructions primer + governance templates	MCP tool: apply_primer / audit_project

Agent rule: Query the data model API before reading source files.

Invoke-RestMethod "http://localhost:8010/model/agent-guide"   # complete protocol
Invoke-RestMethod "http://localhost:8010/model/agent-summary" # all layer counts

---

Note (2026-02-20): PROJECT-39 (39-ado-dashboard) has been scaffolded and moved to its own folder.

• Full plan: 39-ado-dashboard/PLAN.md
• Current status: 39-ado-dashboard/STATUS.md
• Original planning doc (archived): 38-ado-poc/PROJECT-39.md

What Is This?

The EVA ADO Command Center is the orchestration hub for all EVA Platform projects. It reads the live backlog from Azure DevOps, determines which projects have active work items, and dispatches project-specific copilot-agent runners to execute their DPDCA cycle.

This is not a passive mirror of sprint history. It is the control plane for scrum execution across the EVA platform:

• ADO is the single source of truth for what work is active, blocked, or done
• Each project runner is dispatched from here ? it reads its WI from ADO, executes the DPDCA loop, and pushes state back
• Skills for all runners are mastered in 29-foundry and sourced centrally
• The EVA Data Model (37-data-model) is a first-class citizen ? ADO WIs can reference data model entities, and the data model will expose sprint metadata as queryable structured data
• APIM (17-apim, marco-sandbox-apim) injects x-eva-* cost attribution headers on every API call ? tied to the user, business unit, project, client, sprint, and WI tag ? feeding both the EVA User Directory (31-eva-faces) and the FinOps Dashboard (14-az-finops)
• Once eva-brain is APIM-integrated, 39-ado-dashboard delivers the EVA Home page (product tile grid, 23+ products like eva-suite.github.io) and ADO sprint views (/devops/sprint) inside 31-eva-faces, showing live ADO sprint state per project to any authenticated user without needing ADO access
• 14-az-finops is a separate pipeline concern ? Azure Cost Management export ? Power BI analytics consuming the x-eva-* attribution headers

---

Platform Architecture (5 Layers)

??????????????????????????????????????????????????????????????????????????
?  Layer 1: 29-foundry ? Central Agentic Capabilities Hub              ?
?  Master .github/copilot-skills library ? sourced to all projects     ?
????????????????????????????????????????????????????????????????????????
                        ?
????????????????????????????????????????????????????????????????????????
?  Layer 2: 38-ado-poc ? EVA ADO Command Center (this repo)            ?
?  Reads ADO board ? identifies active WIs ? dispatches project runners  ?
?  ADO ? 37-data-model (bidirectional WI ? entity awareness)            ?
??????????????????????????????????????????????????????????????????
             ?                   ?                   ?
      ??????????????    ?????????????    ?????????????
      ? 31-eva-faces?    ?33-brain-v2?    ? 3N-...   ?   Layer 3: Project Runners
      ? DPDCA runner?    ?DPDCA runner?    ? DPDCA    ?   Each has own SESSION-STATE
      ? skills?foundry?  ?skills?foundry?  ? runner   ?   State read from ADO at
      ??????????????    ?????????????    ????????????   session start
             ?                   ?
      ???????????????????????????????
      ?  Layer 4: 17-apim ? API Gateway + Cost Attribution    ?
      ?  eva-brain/roles ? APIM ? eva-faces                   ?
      ?  Injects x-eva-user-id, x-eva-project-id,             ?
      ?           x-eva-business-unit, x-eva-client-id,        ?
      ?           x-eva-sprint, x-eva-wi-tag on every call     ?
      ?  /evaluate-cost-tags (Roles API) ? deployed           ?
      ????????????????????????????????????????????
                          ?              ?
      ???????????????????????????????  ??????????????????????????????
      ?  Layer 5a: 39-ado-dashboard ?  ?  Layer 5b: 14-az-finops    ?
      ?  EVA Home: product tiles    ?  ?  Azure Cost Mgmt export    ?
      ?  23+ products, ADO badges   ?  ?  ? Power BI analytics      ?
      ?  /devops/sprint ADO views   ?  ?  cost per sprint/WI/client ?
      ???????????????????????????????  ??????????????????????????????

Active Projects in ADO

All 18 eva-foundation projects are loaded. Full board: https://dev.azure.com/marcopresta/eva-poc/_boards/board/t/eva-poc%20Team/Epics

Folder	Epic id	Maturity	PBIs	Done
14-az-finops	15	empty	3	0
15-cdc	16	empty	2	0
16-engineered-case-law	17	poc	4	2
17-apim	18	poc	4	2
18-azure-best	19	active	4	2
19-ai-gov	20	poc	4	2
20-AssistMe	21	poc	4	2
24-eva-brain	22	retired	2	2
29-foundry	23	active	4	2
30-ui-bench	24	poc	3	1
31-eva-faces	25	active	4	2
33-eva-brain-v2	26	active	7	6
63-factory-context-auditor	27	idea	2	0
35-agentic-code-fixing	28	poc	2	1
36-red-teaming	29	active	3	1
37-data-model	30	active	7	6
38-ado-poc	31	active	5	2
39-ado-dashboard	32	poc	2	0

Totals: 18 Epics ? 48 Features ? 55 PBIs ? 27 Done / 28 New
Last import: run6 + 37-data-model API import, 2026-02-21, zero errors

---

Folder Structure

38-ado-poc/
??? README.md          ? you are here
??? PLAN.md            ? layered architecture + design decisions
??? ROADMAP.md         ? future: 39-ado-dashboard, foundry skill centralization, finops pipeline
??? ACCEPTANCE.md      ? definition of done for the Command Center itself
??? STATUS.md          ? live board state: 18 projects, 52 PBIs, run6 clean
??? ANNOUNCEMENT.md    ? stakeholder announcement
??? APIS.md            ? ADO REST API reference (every endpoint used)
??? URLS.md            ? board, backlog, query, and WI direct links
??? docs/
?   ??? ADO/
?       ??? ONBOARDING.md        ? schema, script reference, add-project guide
?       ??? THREE-SYSTEM-WIRING.md ? architecture, PR bridge, webhook, deployment
?       ??? idea/                ? idea intake templates
??? scripts/
    ??? ado-import-project.ps1   ? shared import engine (reads ado-artifacts.json)
    ??? ado-onboard-all.ps1      ? orchestrator: batch import all 18 projects
    ??? ado-generate-artifacts.ps1 ? idea intake parser: markdown ? ado-artifacts.json
    ??? logs/                    ? timestamped transcripts (yyyyMMdd-HHmm-*.log)
    ??? ado-setup.ps1            ? legacy: one-time brain v2 / faces population
    ??? ado-bootstrap-pull.ps1   ? reads board ? builds WI queue for runner
    ??? ado-close-wi.ps1         ? pushes WI ? Done with test/coverage metrics
    ??? ado-create-bug.ps1       ? self-improvement: P0/P1 finding ? ADO Bug

# Per-project (inside each project folder):
<project>/ado-artifacts.json     ? ADO import schema for that project
<project>/ado-import.ps1         ? thin wrapper calling ado-import-project.ps1

---

Scripts & Automation

Script	Purpose
scripts/ado-import-project.ps1	Shared import engine ? reads ado-artifacts.json, creates Epic ? Features ? PBIs, assigns sprints, sets Done states
scripts/ado-onboard-all.ps1	Orchestrator ? discovers all project folders, runs import engine for each, logs everything
scripts/ado-generate-artifacts.ps1	Idea intake parser ? converts a structured markdown brief into ado-artifacts.json
<project>/ado-import.ps1	Per-project thin wrapper ? sets paths and calls the shared engine

Full documentation: docs/ADO/ONBOARDING.md
Three-system wiring: docs/ADO/THREE-SYSTEM-WIRING.md

Quick Onboard ? One-Time Full Load

# Dry run (no PAT needed)
cd C:\eva-foundry\eva-foundation\38-ado-poc\scripts
.\ado-onboard-all.ps1 -DryRun

# Live import
$env:ADO_PAT = "<your-pat>"
.\ado-onboard-all.ps1 -RunImport
# Logs saved to: scripts/logs/yyyyMMdd-HHmm-ado-onboard-all-live.log

Add a New Project

# 1. Create ado-artifacts.json in the new project folder
# 2. Create the thin wrapper
.\ado-generate-artifacts.ps1 -SourceMd ..\41-new-project\PLAN.md -OutputDir ..\41-new-project

# 3. Preview
.\41-new-project\ado-import.ps1 -DryRun

# 4. Import
$env:ADO_PAT = "<your-pat>"
.\41-new-project\ado-import.ps1

---

Quick Start ? Dispatch a Runner

# 1. Set PAT (never stored in any file)
$env:ADO_PAT = "<your-pat>"

# 2. Read Command Center board ? see all active WIs across all projects
cd C:\eva-foundry\eva-foundation\33-eva-brain-v2
.\scripts\ado-bootstrap-pull.ps1
# Output: markdown WI table ? this becomes the session context for the runner

# 3. Dispatch a runner to a specific project
# The agent reads the WI from ADO, cd's into the project repo, and runs its DPDCA loop
# using the copilot-skills sourced from 29-foundry

# 4. After DPDCA completes: push result back to ADO
.\scripts\ado-close-wi.ps1 -WiTag "WI-7" -TestCount 600 -Coverage "73" -Notes "APIs deployed"

# 5. Promote a self-improvement finding to a tracked Bug
.\scripts\ado-create-bug.ps1 -Title "axe timeout in waitFor" -Severity "2 - High" -Sprint "Sprint-6"

---

Security Note

The PAT is always passed as $env:ADO_PAT. It is never written to any file in any repo, including .env.ado.
.env.ado contains only IDs and URLs ? safe to commit.

---

Key Relationships

From	To	Nature
38-ado-poc	29-foundry	Sources master skill library for all dispatched runners
38-ado-poc	37-data-model	Bidirectional ? ADO WIs reference data model entities; data model exposes sprint metadata
38-ado-poc	31-eva-faces	Dispatches DPDCA runner; receives WI close events
38-ado-poc	33-eva-brain-v2	Dispatches DPDCA runner; receives WI close events
33-eva-brain-v2 + 17-apim	39-ado-dashboard	APIM route exposes ADO data; EVA Portal + sprint views consume it
39-ado-dashboard	31-eva-faces	Delivered as EVAHomePage.tsx + SprintBoardPage.tsx inside eva-faces