dge-plan-graph

DGE Plan Graph

Use this skill after /dge-design has produced ready requirements.

Purpose

Create the executable delivery graph: tracks, nodes, dependency edges, and validation contracts. This skill plans work; it does not implement code.

Core rules

1. The graph is the source of truth.
2. Every executable node must map to at least one requirement.
3. Every node must include validation before it can be synced or executed.
4. Split large nodes until each node has one clear validation contract.
5. Preserve dependency edges explicitly; do not flatten into a checklist.
6. Prefer the fewest tracks and nodes that cover the must requirements; add a track or node only when a requirement or validation contract demands it. should/could requirements do not need their own track.

Preflight: require the DGE CLI

Before planning, run the shared preflight (one callable place for every dge-* skill and the /dge-deliver conductor). Planning runs after dge init, so the graph must exist and validate:

dge preflight || npx --no-install dge preflight

If it exits non-zero, stop and tell the user to install DGE first (the plugin ships dge on the PATH, or npm install --save-dev github:rafaelolsr/delivery-graph). The dge CLI is the only writer of delivery-graph/graph.json; never hand-write or hand-edit it.

Workflow

1. Load design artifacts

Read delivery-graph/graph.json or the user-provided graph path. Confirm:

• no unresolved blocker gaps
• at least one demand exists
• requirements have acceptance criteria
• requirements have validation methods

Also read prior learnings so the plan does not repeat a known mistake:

dge learnings --about "<keywords from these requirements>" --json   # or: npx dge learnings ...

A learning may change how you split nodes or what a validation contract must assert (e.g. "substring evidence is weak — gate correctness on a live check"). Fold any applicable learning into a node's validation contract.

2. Create tracks

Use tracks to group related nodes. Most demands need only implementation + validation; add others only when a requirement demands them. Common tracks:

• product-contract
• research
• implementation
• integration
• validation
• documentation
• release

3. Create nodes

Each node must include:

- id: NODE-001
  title: "..."
  type: implementation
  track: TRK-implementation
  requirement_ids: [REQ-001]
  depends_on: []
  status: ready
  validation:
    required:
      - "..."
    evidence_path: delivery-graph/evidence/NODE-001/
  sync:
    linear_issue_id: null
    ado_task_id: null

4. Validate graph shape

Check:

• no dependency cycles
• no node depends on itself
• every dependency points to an existing node
• every requirement is covered by at least one node
• every node has at least one validation requirement

5. Save graph

Create tracks and nodes through the CLI (dge add-track, dge add-node); it updates delivery-graph/graph.json for you. Do not write graph.json yourself.

Output

Follow the shared output convention (see skills/README.md): lead with a bold one-line synthesis (e.g. "Graph planned: N tracks, M nodes, K ready"), then the demand's progress indicator (see skills/README.md), then the detail:

• graph id
• tracks created
• nodes created
• ready nodes
• blocked nodes
• requirements without nodes, if any

Then a ## Next block: /dge-sync or /dge-work-node NODE-###.

CLI contract

The DGE CLI is required (see Preflight) and is the only writer of the canonical store. Use dge if it is on the PATH (the plugin ships it), otherwise npx dge:

dge add-track ...        # or: npx dge add-track ...
dge add-node ...
dge status

Never edit graph.json by hand. If the CLI is unavailable, stop and install it (Preflight).