The Task Analysis editor now works the way analysts expect. Add Step inserts a sibling directly under the currently selected step rather than jumping to the bottom of the list, a new Sub Step button creates child steps in place (step 3 becomes 3.1), and whole rows can be selected rather than individual cells. The top toolbar is now sticky so its controls survive scrolling through a long decomposition, columns are drag-resizable, and the error indicator has moved from red to a neutral colour so it is no longer mistaken for a Safety-Critical flag.

A new tab strip under the editor toolbar lets analysts carve a 300-plus step decomposition into named groups, replacing the spreadsheet-tab workflow many teams fell back on. An "All" tab is always present, and a half "+" tab defines a new grouping (for example steps 3 to 5); selecting a group filters the view to just those steps without disturbing the underlying step order or numbering. Groupings are saved with the document so they survive reload.

The SCTA home and session pages now describe themselves in the analyst's own terms - Major Accident Hazard, Site and Task - in place of the generic platform identifiers, with a linked-context strip showing what each analysis is grounded in. A duplicated information page has been removed, the spawn control is now state-aware (Open SCTA when one is linked, Spawn SCTA when none is), and an Open Registers / Open Task button returns the user to the right place with the correct navigation highlight, ending the stale-menu confusion reported in use.

Task criticality bands have been recalibrated to OTO 1999/092 production-task ranking - 0 to 3 Low, 4 to 8 Medium, 9 and above High - with emergency scenarios automatically classified High and the raw score left uncapped. The per-task identifier (for example TR-006) now appears as a leading column, the screening detail sits beside the line-item details rather than below them, editing a Major Accident Hazard defaults to all sites, and SCTA studies can be linked directly from the register.

An error can now hold more than one classification at once - for example both a Slip and a Mistake - and the row pills update to reflect what has actually been entered. Action capture has moved out of the PIF popup into the dedicated "+ Action" flow, the unjustified Post-Mitigation Risk box and "Mistake" recommendation source have been removed, and the analysis surface now accepts data in place instead of visibly reloading from the bottom of the page on every edit.

A new licence panel in Settings > About shows the licence currently in use and lets users change or upgrade their key without leaving the application. Moving from a trial to a commercial licence now happens in place: enter the new key and the licence updates immediately, with no manual deactivation step required.

A full review of button styling across Charlie, Harry and the newer Classic theme has tightened contrast wherever text or controls risked fading into their surface. Pastel rating buttons, RULA and REBA launch controls, and report-page hover states now meet contrast expectations in every theme, and selected list items render legibly in Classic instead of appearing as white-on-light text.

Tool records are now explicitly scoped to their own project on save, closing a latent path where data from one project could be carried into another when two projects were opened in a single session. The reliability documents - SHERPA, THERP, SPAR-H, uFMEA and URRA - now stamp their project identity correctly, including older files that are brought up to date as they load, so no document is dropped or misfiled on re-save.

Deleting a newly created error on the Task Analysis page now also removes it from the Human Error Analysis, while authored analysis is protected by a confirm rather than silently lost. Multi-line barrier evaluation text now persists on click-away instead of collapsing to a single row, and the "Source: ErgoBow Barrier" label no longer appears when there is no barrier linked.

Fixed a selection defect in the registers where a row could stay stuck selected, and two rows could end up selected at once. Selection is now reliably single-row across the register views.

The automated user-interface test suite introduced in 1.3 continues to grow, exercising more tools through their actual interface before each build. Coverage widens release on release as part of the standard pre-release gate alongside the existing unit and integration tests.

Every change in this release - the new SCTA groupings, multi-select classification and project-scoped saves - is additive and backward compatible. Existing assessments and older project files continue to open cleanly, with legacy data brought up to date on load and no migration step required.

Behind the scenes, we are building the infrastructure for AI-assisted support inside ErgoSphere. The entitlement, licensing and service foundations are being put in place now so the feature can be switched on cleanly when it is ready. There is nothing to enable in this release - this is an early heads-up so practitioners and teams can anticipate what is coming. More detail will follow in a future update.