teamcraft-glgd:define-roadmap

---
name: teamcraft-glgd:define-roadmap
description: Produce a multi-sprint product roadmap in Google Drive — the planning layer between a baselined PRD and individual sprint planning. Organises what gets built in what order across sprints, with priorities, dependencies, and capacity rationale. Feeds directly into plan-sprint.
argument-hint: "[Drive PRD URL or file ID — optional]"
disable-model-invocation: true
user-invocable: true
allowed-tools:
  - mcp__gitlab__list_projects
  - mcp__gitlab__list_milestones
  - mcp__google-drive__list_accounts
  - mcp__google-drive__search_files
  - mcp__google-drive__download_file
  - mcp__google-drive__create_file
  - mcp__google-drive__update_file
---

## Goal

Produce a roadmap that answers: given everything in the PRD, what gets built in what order, over how many sprints, and why in that sequence? The roadmap is the bridge between "here are all the requirements" and "here is what we commit to this sprint." It makes sequencing decisions explicit and visible so sprint planning is selection from a ranked backlog, not improvisation.

The roadmap is an AI-native artifact — it must be structured so an agent loading it for sprint planning can immediately understand priorities, dependencies, and sequencing rationale without human interpretation.

## Hard Constraints

- The roadmap organises what is already in the PRD. It does not introduce new requirements. If the user wants to add scope, direct them to update the PRD first.
- Sequencing must be grounded in reasons: dependencies between capabilities, risk reduction, walking skeleton first, client contractual milestones, technical prerequisites. Arbitrary sequencing is not a roadmap.
- If a Drive file operation fails with a path error, read the error message to identify a valid accessible host path and retry with it.

## Resolve Drive Account

Call `mcp__google-drive__list_accounts` before any other Drive operation:
- **No accounts** — Drive is not configured. Tell the user and skip Drive operations.
- **One account** — Use it. Pass `account_email` explicitly on every Drive tool call this session.
- **Multiple accounts** — Present the list, ask which account to use.

## Load Inputs

If `$ARGUMENTS` contains a Drive URL or file ID, load the PRD directly. Otherwise ask the user if they can point at the PRD before searching.

Ask whether a GitLab project exists — if so, load sprint history from `mcp__gitlab__list_milestones` to understand what has already shipped and give the roadmap accurate starting context.

Ask whether tech decisions are available in Drive. Load if the user can point at them — sequencing often depends on technical prerequisites.

## Produce the Roadmap

Work conversationally. The PRD contains all requirements — the task is to sequence and group them into cohesive sprint-sized chunks with clear rationale.

For each roadmap item, establish: what it is (capability or feature area from the PRD), why it comes at this point in the sequence (dependency, risk, value), roughly how large it is (sprint count estimate), and what it unlocks next.

See `references/example-roadmap.md` for the expected structure and depth.

Surface sequencing decisions that need the user's input — when two capabilities have no dependency between them, the PM chooses priority. Do not assume.

## Review and Store

Show the complete roadmap for review. Do not store until the user confirms.

Ask the user where in Google Drive to store it — which folder. Never save to the Drive root. Record the URL.

## Done

Share the Drive URL. Add the roadmap URL to `.teamcraft/project.md` alongside the PRD and tech decisions URLs — plan-sprint will load it as context for sprint selection. The natural next step is plan-sprint: select the first sprint's issues from the roadmap's first phase.