# Mission: Deciding the fate of Sprints & Agile in LVC WorkOS PM

## Why
The `/pm` module ships a "Sprints & Agile" surface that is backed by a table
nobody writes to, sitting next to a `/tasks` board that *is* real and is wired
into timesheets, leave, capacity and performance goals. Chakresh needs to make
an architecture call — merge the two surfaces, wire the sprint up properly, or
cut it — and stop the module's documentation from describing a schema that does
not exist.

## Success looks like
- Can state in one sentence what a sprint is *allowed to be* in this schema, and
  why that makes it a timebox rather than a work item.
- Can point at the three files that make `/pm/sprints` non-functional today, and
  say which of the three is the real bug.
- Has a written verdict on merge-vs-separate for `pm_sprints` and `/tasks`,
  justified from the schema rather than from taste.
- Has deleted the fiction from `PM_MODULE_GUIDE.md` §5 (14 `pm_*` tables, 10 of
  which do not exist) so the next engineer is not misled.

## Constraints
- Founder/builder of the platform: deep TypeScript and Postgres fluency. Do not
  teach React, Next.js, or SQL syntax. Teach the *domain judgement* that the
  code cannot tell them.
- The decision must survive contact with the rest of the platform: sprints
  cannot be designed in isolation from `capacity.service.ts`, `timesheet_entries`,
  `leave_requests`, and `performance.goals`.
- Verdict should be small enough to ship. Not a rewrite of the PM module.

## Out of scope
- The `Action Required` paradigm — real problem, separate decision, separate
  lesson. It is noted here only because it is adjacent and currently faked.
- Building the workspace/portfolio/program 10-level tree from `PM_MODULE_GUIDE.md`.
  Treat as aspirational design, not as a backlog.
- Kanban UI ergonomics, Gantt rendering, RAID matrix behaviour.