Ralph Bridge

Generate `prd.json` from BMAD V3 planning artifacts.

This command reads the V3 epic/story definitions and sprint plan, then outputs a machine-readable `prd.json` at the project root containing all stories with dependency graph and gate nodes.

**IMPORTANT — Windows encoding**: Always read/write prd.json with `encoding='utf-8'`. The `gates` field is a **list** of gate objects, `sprints` is an **integer** (sprint count). Derive `totalStories` and `totalStoryPoints` dynamically from the stories array — never hardcode.

## Steps

### 1. Read Source Artifacts

Read these files to extract story definitions:
- `_bmad-output/project-planning-artifacts/acm-ai/05-epics-and-stories.md` — All story definitions
- `_bmad-output/project-planning-artifacts/acm-ai/04-architecture.md` — Architecture context
- `docs/sprint-artifacts/v3-sprint-plan.md` — Sprint assignments and schedule
- `docs/sprint-artifacts/sprint-status.yaml` — Current story status (if any already done)

### 2. Build Story Objects

For each of the 33 V3 stories (E30-S1 through E34-S4), create a JSON object:

```json
{
  "id": "E30-S1",
  "title": "Salesforce Schema Config Loader",
  "epic": "E30",
  "sprint": "V3-1",
  "storyPoints": 5,
  "riskLevel": "HIGH|MEDIUM|LOW",
  "storyType": "backend|frontend|both",
  "acceptanceCriteria": ["AC1: ...", "AC2: ..."],
  "passes": false,
  "notes": "",
  "model": "sonnet",
  "complexity": "complex|moderate|simple",
  "priority": "P0|P1|P2",
  "dependencies": [],
  "keyFiles": ["open_notebook/extractors/parsers/config_loader.py"],
  "techSpecFile": null,
  "implementedDate": null
}

3. Apply Dependency Map

Use this exact dependency graph for all 33 stories:

E30 — Foundation:

• E30-S1: [] (no deps — entry point)
• E30-S2: ["E30-S1"]
• E30-S3: ["E30-S1"]
• E30-S4: ["E30-S1"]
• E30-S5: ["E30-S2", "E30-S3"]
• E30-S6: ["E30-S3"]
• E30-S7: ["GATE:SCHEMA_FREEZE", "E31-S2"]
• E30-S8: ["E30-S1"]

E31 — Extraction:

• E31-S1: ["GATE:SCHEMA_FREEZE"]
• E31-S2: ["E31-S1"]
• E31-S3: ["E31-S2"]
• E31-S4: ["E31-S2"]
• E31-S5: ["E31-S3", "E31-S4"]
• E31-S6: ["E31-S5"]
• E31-S7: ["E31-S5"]

E32 — AI Processing:

• E32-S1: ["E31-S5", "E30-S7"]
• E32-S2: ["E32-S1"]
• E32-S3: ["E32-S2", "E30-S4"]
• E32-S4: ["E30-S6"]
• E32-S5: ["E32-S1", "E32-S2", "E32-S3", "E32-S4"]
• E32-S6: ["E30-S8"]

E33 — Frontend:

• E33-S1: ["GATE:SCHEMA_FREEZE", "E31-S7"]
• E33-S2: ["GATE:SCHEMA_FREEZE", "E30-S2"]
• E33-S3: ["GATE:AI_COMPLETE", "E30-S4"]
• E33-S4: ["E33-S2", "E33-S3", "E30-S4"]
• E33-S5: ["E31-S4", "E33-S2"]
• E33-S6: ["E31-S4", "E33-S2"]
• E33-S7: ["E30-S2", "E33-S2"]
• E33-S8: ["E30-S2", "E30-S3", "E33-S4"]

E34 — Integration:

• E34-S1: ["E31-S7", "E33-S2"]
• E34-S2: ["E33-S2", "E33-S4", "E31-S7"]
• E34-S3: ["E31-S5", "E32-S5"]
• E34-S4: ["GATE:UI_COMPLETE"]

4. Apply Risk & Complexity

HIGH risk stories (need architect guidance):

• E30-S1 (first V3 story, defines patterns), E30-S2 (building model, FK design), E30-S4 (36 combos + 114 mappings)
• E31-S3 (consensus layer core), E32-S3 (AI correction loop), E33-S2 (two-view AG Grid)

5. Apply Story Types

• backend: E30-S1, S3, S4, S5, S6, S7, S8, E31-S1..S7, E32-S1..S6
• frontend: E33-S1..S8
• both: E30-S2 (building model + API + frontend list), E34-S1..S4

6. Build Gates Array

{
  "gates": [
    {
      "id": "GATE:SCHEMA_FREEZE",
      "name": "Schema Freeze",
      "triggerStory": "E30-S6",
      "unlocked": false,
      "blocksEpics": ["E31", "E32", "E33", "E34"],
      "description": "All SF schema definitions finalized. BAR->SF vocabulary locked."
    },
    {
      "id": "GATE:EXTRACTION_COMPLETE",
      "name": "Extraction Complete",
      "triggerStory": "E31-S6",
      "unlocked": false,
      "blocksEpics": [],
      "description": "Multi-provider extraction pipeline operational with consensus layer."
    },
    {
      "id": "GATE:AI_COMPLETE",
      "name": "AI Complete",
      "triggerStory": "E32-S5",
      "unlocked": false,
      "blocksEpics": [],
      "description": "AI validation, correction, and batching pipeline operational."
    },
    {
      "id": "GATE:UI_COMPLETE",
      "name": "UI Complete",
      "triggerStory": "E33-S8",
      "unlocked": false,
      "blocksEpics": [],
      "description": "All frontend views operational with AG Grid, provenance UI, and dashboards."
    }
  ]
}

7. Write prd.json

Write the complete JSON to prd.json (with encoding='utf-8') at the project root with this top-level structure:

{
  "version": "3.0",
  "project": "ACM-AI V3",
  "generatedAt": "<ISO timestamp>",
  "totalStories": "<len(stories)>",
  "totalStoryPoints": "<sum of all storyPoints>",
  "sprints": "<number of distinct sprints>",
  "stories": [ ... ],
  "gates": [ ... ],
  "metadata": {
    "source": "ralph-bridge",
    "artifacts": [
      "05-epics-and-stories.md",
      "04-architecture.md",
      "v3-sprint-plan.md"
    ]
  }
}

Note: totalStories, totalStoryPoints, and sprints must be computed dynamically from the stories array — never hardcode. sprints is an integer (count of distinct sprint values). gates is a list.

8. Validate

After writing, verify:

• Story count matches totalStories
• All gates referenced in story dependencies exist in gates list
• E30-S1 has dependencies: []
• No circular dependencies
• All story IDs are unique
• Sprint assignments cover all stories

Report the result: total stories, total SP, gates, and first eligible story (should be E30-S1). ```