Skip to content

Build backlog

Generated from the architecture model

This page is rendered from model/**/*.yaml by tools/docs_model during mkdocs build. Edit the YAML, not this page; hand-written design lives on the linked service, platform and plane pages.

A design artefact, not a tracker

This backlog is generated from model/backlog/ and validated against the architecture model. No work item exists in any tracker, and no item carries a date, an estimate or an assignee (ADR-0029).

The backlog has 9 increments, 20 epics, 185 features, 823 PBIs and 2886 tasks. Epics follow the structure of the system: one per platform, plus readiness, component extensions, Factory assets and outcome verification. Every feature carries the increment whose outcome needs it, so the same items read two ways: by increment and by epic.

Only the first wave is elaborated into PBIs and tasks (rolling wave): I0 Readiness and I1 Generate a microservice and I2 Deliver it and I3 Studio and I4 Trust and accountability. Later increments stay at epics and features until they are next.

Increments

Order is by dependency, not by date.

Increment Outcome Depends on Features PBIs Tasks
I0 Readiness Every repository of the first outcome scaffolds from its template and builds, and one call chain carries a trusted tenant and project scope from the edge to a service. — 27 19 140
I1 Generate a microservice A vision submitted through the API or the command-line client ends as an approved change proposal that contains a generated, built and tested microservice (decision D65); the run is audited from its first command and stops at its budget cap (decision D80). I0 48 6 347
I2 Deliver it The merged service is built by a pipeline, deployed to a preview environment and promoted under approval; the infrastructure it runs on is planned by the Factory from the design. I1 21 3 126
I3 Studio The same chain is submitted, watched and decided in Factory Studio. I1 17 3 85
I4 Trust and accountability A release publishes only with evaluation evidence, approvals follow the rules of the project, and a run recovers after a worker is killed. I1 20 1 125
I5 Knowledge and connections A run on an existing repository uses scoped retrieval with citations, and external systems connect through governed connectors. I4 15 — —
I6 Full product generation A multi-module product with a user interface, a gateway and shared libraries is generated from a vision. I2, I5 21 5 —
I7 Operate and commercialise Generated products are inventoried, upgraded and off-boarded, and Factory usage is billed by edition. I6 10 4 —
I8 Expansion Marketplace distribution, self-hosted models and agent-to-agent delegation. I7 6 — —

Epics

Epic Features By increment
Control Plane (PF-01)
AIFACTORY-EPIC-001
10 I1: 6 I2: 1 I4: 1 I7: 1 I8: 1
Agent Mesh (PF-02)
AIFACTORY-EPIC-002
3 I1: 2 I3: 1
Model & Compute Platform (PF-03)
AIFACTORY-EPIC-003
3 I1: 2 I8: 1
Knowledge Platform (PF-04)
AIFACTORY-EPIC-004
11 I1: 6 I5: 5
Tools & MCP Platform (PF-05)
AIFACTORY-EPIC-005
4 I1: 3 I5: 1
Integration Platform (PF-06)
AIFACTORY-EPIC-006
7 I1: 3 I2: 1 I5: 3
Marketplace (PF-07)
AIFACTORY-EPIC-007
3 I8: 3
DevOps / GitOps (PF-08)
AIFACTORY-EPIC-008
5 I1: 1 I2: 4
Runtime Cloud (PF-09)
AIFACTORY-EPIC-009
5 I1: 1 I2: 4
Product Fleet Operations (PF-10)
AIFACTORY-EPIC-010
4 I7: 4
Data & Event Platform (PF-11)
AIFACTORY-EPIC-011
4 I4: 4
Governance (PF-12)
AIFACTORY-EPIC-012
5 I1: 4 I4: 1
Observability (PF-13)
AIFACTORY-EPIC-013
7 I1: 1 I3: 1 I4: 5
Factory Studio (PF-14)
AIFACTORY-EPIC-014
10 I3: 10
Template & Library (PF-15)
AIFACTORY-EPIC-015
4 I1: 4
Generated SaaS (PF-16)
AIFACTORY-EPIC-016
2 I6: 2
Engineering readiness
AIFACTORY-EPIC-017
9 I0: 7 I1: 2
Ecosystem component extensions
AIFACTORY-EPIC-018
44 I0: 19 I1: 6 I2: 3 I3: 4 I4: 2 I5: 1 I6: 5 I7: 4
Factory assets: workflow definitions, agents and skills
AIFACTORY-EPIC-019
36 I1: 6 I2: 7 I4: 6 I5: 4 I6: 13
Outcome verification
AIFACTORY-EPIC-020
9 I0: 1 I1: 1 I2: 1 I3: 1 I4: 1 I5: 1 I6: 1 I7: 1 I8: 1

Definition of Ready

  • The item has its id, parent, sequence and a description that can be read without the rest of the backlog.
  • A PBI has two to five numbered acceptance criteria, each citing a traced id or a source path; a task has a kind and an expected output.
  • A PBI traces the endpoints, events, steps and scenarios it delivers, depends only on PBIs of its feature or on features and epics, and carries at least one verification task.
  • Every model reference resolves in the architecture model of this repository.
  • Tenancy, security, data, observability, localization and user-experience impact is handled or marked not applicable.
  • Dependencies on other features are delivered in the same or an earlier increment, or a stub is declared with the increment that completes it.
  • Owner decisions the item depends on are recorded in the decision log.

Other views and exports

  • Scenarios: the derived test-scenario catalog every PBI traces (one per endpoint, event, workflow step, condition outcome and gateway route class), with the ratchet state.
  • Build order: the derived order of all features, templates and libraries first, then the Factory code by increment, carried into the exports as a Stack Rank.
  • Coverage: where every service, bounded context, workflow definition, requirement, work area and component is placed.
  • Dependencies: the order of increments and the dependencies between features.
  • Draft work items, JSON and Azure DevOps import, CSV, with one CSV per increment, one per epic under export/epics/ and packed files of at most 1,000 rows (aifactory-backlog-ado-NN.csv) beside them. They are drafts: createdInTracker is false, and nothing is imported before the owner approves.

How an item is identified

An id such as AIFACTORY-FEAT-001 is assigned once and never renumbered or reused; the ledger is model/backlog/ids.lock.yaml. The sequence number shown in the exports is derived from the order of the files and may change when an item is inserted.