Progress reports from derived data

The principle

Progress is never entered by a human or a tool — it's computed from task counts. Complete tasks and the numbers roll up automatically: task → module → milestone → project. A status report is therefore just a read, never a write.

The read

curl "https://pm.blarass.com/v1/progress?project_id=<project-uuid>" \
  -H "Authorization: Bearer <key>"

Or from an agent:

{ "tool": "get_project_progress", "arguments": { "project_id": "…" } }

The dated view (timeline)

curl "https://pm.blarass.com/v1/roadmap?project_id=<project-uuid>&from=2026-09-01&to=2026-10-31" \
  -H "Authorization: Bearer <key>"

Milestones ordered by target_date (undated last), each with total_tasks / done_tasks / progress_pct — a ready-made status table for stakeholders.

Recipe: weekly status update

1. Pull /v1/progress?project_id=… — the roll-up.
2. Pull /v1/tasks?project_id=…&status=in_progress — what's moving.
3. Pull /v1/tasks?project_id=…&assignee=me if reporting narrowly.
4. Compose the report. If your coding agent is writing the update, it has all three endpoints over MCP already — zero extra integration.

Related: Blockers · Comments workflow