Keep the workflow that fits your current stage. Each accepted protocol gets a named version, its parent version, a dated release record and a rollback path. Saved reports keep the rule version that applied when they were recorded.
| Version | Purpose and status | Open |
|---|---|---|
| Cockpit eight v1 · live.1 | Current presentation Accepted compact decision meter, evidence-based messages and larger 26px preparation cards. Eight-decision policy starts no earlier than September 18, 2026, 4:00 p.m. America/Phoenix, after the previous session is closed, flat and fresh preparation is completed. Ongoing sessions retain existing policy until transition. | Live cockpit Isolated test cockpit |
| Behavioral pressure v0.3.1 | Previous live reference Advisor reviewed. Approved state and deviation weights, bounded recovery, four categories, separate live evidence store. 78 unit tests and both route integration suites passed. AI interpretation unavailable. | Live cockpit Isolated test cockpit |
| Live baseline · 2026-09-16 UI 8d4cffbdc81d | Archived baseline Existing six-card workflow and safeguards. Frozen rollback snapshot retained. | Live cockpit Baseline rehearsal |
| Behavioral 1.0.0-test.1 Parent: live baseline | Test only · awaiting acceptance Original compact meter. Superseded visually by UI 1.1; preserved in the release archive and guarded rollback backups. | Behavioral test cockpit |
| Behavioral UI 1.1 Parent: behavioral-1.0.0-test.1 | Previous visual version Large color-coded Performance / Emotion curve, 0–100 illustrative scale, reported activation band and separate execution evidence. Rules remain 1.0.0-test.1. Browser walkthrough pending. | UI 1.1 test cockpit |
| Behavioral UI 1.2 Parent: UI 1.1 | Previous visual version Half-height graph, latest exit-state dot and vertical guide; separate Error section and DCA label. Engine rules unchanged. A combined 0–100 behavioral ranking remains undefined. | UI 1.2 test cockpit |
| Behavioral UI 1.3 Parent: UI 1.2 | Previous visual version Compact sticky horizontal instrument panel. Curve beside readouts; details expandable. Test label stays visible; safety overlays retain priority. Rules unchanged. | UI 1.3 test cockpit |
| Behavioral pressure v0.1-test.1 Parent: UI 1.3 | Previous test scoring release Weighted global index, critical floors, qualified reset reports and impulse deduplication. Graph-led sticky header and compact colored labels. 30 scorer/reset fixtures and isolated route regression passed. Browser acceptance pending; AI unavailable. | Pressure test cockpit |
| Behavioral UI 1.4 Parent: pressure v0.1-test.1 | Previous test version Graphical report buttons with distinct icons, family colors and explicit selection checkmarks. Pressure weights and recovery policy unchanged. Syntax and UI checks passed; browser acceptance pending. | Current test cockpit |
| Behavioral pressure v0.2-test.1 Parent: UI 1.4 / pressure v0.1 | Previous test version Three compliant decisions may release four burden points; bounded trade/reset recovery. Twelve state choices, red-outline Gamble/Tilt reports. New label weights pending definition. 63 unit checks and real-route integration passed. | Current test cockpit |
| Behavioral pressure v0.3-test.1 Parent: v0.2 | Prior test release All five new labels now weighted. Tilt-state awareness stays distinct from an acted Tilt trade. Gamble/Tilt save immediately; critical floors and bounded recovery retained. 72 unit checks and save/retry checks passed. | Current test cockpit |
The accepted live release has frozen assets and engine modules. The permanent test link uses isolated practice records and the current test branch; experiments are promoted only after review. The baseline rehearsal preserves the original interface. Experimental protocols branch from a recorded baseline; they are not automatically promoted. After review, an accepted release is frozen before a successor is created.
Turning off a new display or AI component does not erase annotations, rewrite grades, change broker settings or unlock a session. A full rollback is performed through the guarded deployment process.
Practice sessions are separate. Each test link has its own store. Resetting the behavioral test archives that practice session and starts a new one. Live trading records stay separate.