99-test-project -- EVA Factory Test Vehicle

Purpose: Test vehicle for EVA Factory script generalization and pattern validation
Owner: 07-foundation-layer (Workspace PM/Scrum Master)
Status: Active -- Testing ground for workspace-level patterns
Last Updated: March 3, 2026
Related: 07-foundation-layer (template source), 51-ACA (pattern source)

---

Overview

99-Test-Project is the Official Test Vehicle for EVA Factory Scripts -- validates that generalized scripts from 07-foundation-layer work correctly before deployment to active projects.

This project validates that:

• seed-from-plan.py works for any project prefix (not just ACA)
• reflect-ids.py annotates PLAN.md correctly
• gen-sprint-manifest.py generates parseable manifests
• All three scripts are idempotent (safe to re-run)
• PROJECT_ID auto-detection works (folder name -> TEST)

Quick Start

# Copy generalized scripts from 07-foundation-layer
cp C:\AICOE\eva-foundry\07-foundation-layer\scripts\*.py C:\AICOE\eva-foundry\99-test-project\scripts\

# Run annotation (creates [TEST-NN-NNN] style IDs in PLAN.md)
cd C:\AICOE\eva-foundry\99-test-project
python scripts\reflect-ids.py --project-id TEST

# Verify veritas-plan.json was created
cat .eva\veritas-plan.json

# Generate sprint manifest
python scripts\gen-sprint-manifest.py --sprint-id "SPRINT-Phase1" --sprint-name "Script Validation"

# Verify manifest
cat .github\sprints\sprint-phase1.md

Scripts

• scripts/seed-from-plan.py -- Parse PLAN.md, generate veritas-plan.json
• scripts/reflect-ids.py -- Annotate PLAN.md with canonical IDs
• scripts/gen-sprint-manifest.py -- Generate sprint manifests

Test Phases

1. Annotation (reflect-ids.py):

   • Read original PLAN.md (has Story N.M.K format)
   • Annotate with [TEST-NN-NNN]
   • Verify idempotent (run twice, second run is no-op)

2. Seeding (seed-from-plan.py):

   • Parse annotated PLAN.md
   • Create .eva/veritas-plan.json
   • Verify 12 stories total (3 epics x 4 stories per epic)

3. Sprint Generation (gen-sprint-manifest.py):

   • Load veritas-plan.json
   • Generate sprint manifest with model assignments
   • Verify JSON block is valid and parseable

Success Criteria (Phase 1, Task 1.3 - Testing)

• All three scripts run without errors on 99-test-project
• PLAN.md annotation produces [TEST-NN-NNN] style IDs
• veritas-plan.json contains 12 stories with correct structure
• Sprint manifest markdown is valid and parseable
• PROJECT_ID auto-detection works (infers TEST from folder name)
• All scripts are idempotent (second run produces same output)
• Data model seeding works (POST to HTTP endpoint)
• Model assignments are correct (gpt-4o-mini for S, gpt-4o for M/L)

Notes

• This project uses PROJECT_ID="TEST" (auto-detected from folder name "99-test-project")
• PLAN.md contains 3 epics x 4 stories = 12 stories total
• All scripts use ASCII-only output (no emoji, no Unicode)
• Scripts are idempotent: safe to re-run multiple times