State Model Round 1 - AFPS Tracker

Stage-zero interrogation for the flow-anchored logical model that will sit between the approved AFPS Tracker user-flow map and UX variation work.

Table Of Contents

Context Snapshot

6 Approved user-flow branches in `design/afps-tracker/flow-tree-afps-tracker.yaml`.
0 Existing state-model cursor files found before this round.
Logical The model covers entities, states, commands, events, read models, policies, and contracts only.
Read-first Source-native AFPS artifacts remain canonical truth; tracker write-back is outside v1 scope.

Primary flow nodes in play: `uf-orient-portfolio`, `uf-verify-selected-path`, `uf-inspect-source-provenance`, `uf-recover-diagnostics`, `uf-handoff-resume`, and `uf-explain-boundaries`.

Assumptions Manifest

Confirm, correct, or flag these source-backed assumptions. Corrections seed the next round or the scope checkpoint.

AssumptionSourceYour decisionCorrection or flag notes
AFPS Tracker models product-path workflow state, not generic project-management tasks. [from positioning]
The canonical source of product-path truth remains AFPS files such as `research/.progress.yaml`, scoped research docs, alignment pages, and archives. [from user-flow]
The core stateful subjects are product paths, source artifacts, evidence links, provenance records, diagnostics, and handoff summaries. [inferred from flow]
Path states must include active, archived, deferred, revisit-candidate, promoted, out-of-scope, unresolved, warning, blocked, and partial interpretations. [from flow-tree and user-flow]
Diagnostics should degrade trust at source, claim, path, and handoff levels rather than hiding partial source-backed value. [from user-flow]
Write-back, storage choices, endpoint shapes, auth, migrations, indexes, and deployment decisions must be deferred to later specification work. [from state-model constraints]
Promoted development/execution work remains outside AFPS Tracker and should hand off toward the separate `exec-loop-tracker` concept when source state supports it. [from journey and user-flow]

Detected Framework Direction

The first-pass detection points to a chunked multi-framework model because the approved flow is stateful, provenance-heavy, diagnostic, and contract-rich.

FrameworkDefault stanceWhy
event-modelingLeadBest fit for ordered scan, resolve, inspect, validate, and handoff events across the flow.
domain-modelAlways includeNeeded to define product paths, evidence refs, diagnostics, provenance records, and handoff summaries as stable logical objects.
state-machineEarlyPath, source, diagnostic, provenance, and handoff lifecycles carry trust-critical status transitions.
data-modelAfter conceptual passesNeeded for the logical ERD and relationship/cardinality clarity, without choosing physical storage.
contract-mapLastNeeded to stabilize command/query request, response, and error shapes after entities and states settle.
event-stormingOptionalUseful only if AFPS vocabulary remains unsettled after this round; current research vocabulary is fairly mature.

Open Questions

Framework Set And Run Order

Should the default framework order stand, or should any framework be added, removed, or reordered?

Recommended: use `event-modeling`, `domain-model`, `state-machine`, `data-model`, and `contract-map` in that order; skip `event-storming` unless you want a vocabulary reset.

Agent confidence: medium

Domain Vocabulary

Which terms should the model preserve exactly, and which terms feel wrong or ambiguous?

Recommended: preserve `product path`, `active path`, `evidence ref`, `alignment provenance`, `diagnostic`, `handoff summary`, `promoted boundary`, and `source-backed claim`; scrutinize `parser confidence` versus `source confidence`.

Agent confidence: medium

Stateful Subjects And Lifecycles

Are there stateful subjects or lifecycle states missing from the default model scope?

Recommended: model lifecycles for product path, source artifact, evidence link, provenance record, diagnostic issue, and handoff summary. Treat repository scan as an event/read-model process, not the central aggregate.

Agent confidence: medium

Command And Query Boundary

Which user/system actions must become logical command/query contracts for UX variations to rely on?

Recommended: include scan portfolio, select product path, inspect evidence, inspect provenance, evaluate diagnostics, generate handoff, copy handoff, and explain boundary as logical contracts.

Agent confidence: medium

Physical Concerns To Defer

What implementation concerns are tempting but should be explicitly deferred out of this logical model?

Recommended: defer database/storage choices, file watcher mechanics, endpoint URLs, auth/permissions, migrations, indexing/search implementation, sandbox details, and deployment.

Agent confidence: high

Riskiest Modeling Assumption

What model assumption would most damage UX variation work if it is wrong?

Recommended: the riskiest assumption is that product path is the main aggregate root; if diagnostics or source artifacts are actually primary, the model and UX branches may organize around the wrong object.

Agent confidence: low

Compile Responses

Answer what you want to change, then compile and paste the YAML into a fresh or current agent session.