A complete new tool for Cognitive Work Analysis supporting all five CWA phases: Work Domain Analysis (Abstraction Hierarchy), Control Task Analysis (Decision Ladders), Strategies Analysis, Social Organisation and Cooperation Analysis (SOCA Matrix), and Worker Competencies (SRK Classification). Each phase features an interactive Cytoscape.js-powered diagram editor with cross-phase data synchronisation, phase completion tracking, and full PDF export. Three richly populated sample studies are included - Stuxnet (Natanz nuclear facility), Sydney Trains (Central-Hornsby corridor), and ATC Terminal Control Area - each demonstrating all five phases with realistic domain content.

The Risk Analytics dashboard now shows risk scores separated by domain (Physical, Cognitive, Reliability, Usability, Systems) with horizontal bar charts, a coverage ring showing assessed versus total domains, and per-domain trend lines. Scores from validated cardinal tools are clearly distinguished from heuristic estimates. Each tool now carries metadata declaring its score domain and capability classification, enabling the dashboard to correctly aggregate only comparable instruments.

A new panel surfaces findings from qualitative tools (HTA safety-critical task analysis, ErgoTrace, ErgoStrata) that cannot produce numeric risk scores, ensuring these tools are visible in project analytics rather than silently excluded from the dashboard.

All tool home pages now feature an Outlook-style details preview pane below the assessment list, separated by a draggable splitter. Selecting an assessment shows editable metadata (area, assessor, notes), contributors, HTA links, issues, requirements, and snapshots in a four-column layout with inline action buttons. Relational management has moved from the right-click menu to this pane for a more accessible workflow.

Ctrl+Shift+Z now works as an alternative redo shortcut alongside Ctrl+Y. Pressing F2 on the tool home page triggers rename for the selected assessment. Delete key focus guards added to six tool pages (HTA, URRA, uFMEA, THERP, SPAR-H, SHERPA) to prevent accidental item deletion when typing in a text field.

The Risk Analytics page has been completely redesigned with domain bar charts, an assessment coverage ring, a three-column domain heatmap replacing the old cross-domain bleed heatmap, and per-domain trend lines replacing the single aggregated trend. Intervention Effectiveness tracking has been moved from Data Explorer to Risk Analytics where it contextually belongs alongside domain scores.

The HTA editor's Overall Risk Assessment panel now shows a count of nodes by SCTA residual risk classification (Unacceptable, Tolerable, Acceptable, Not Assessed) instead of a misleading cross-domain numeric average.

ErgoGaze now contributes a gaze-based risk score to the analytics dashboard. The existing risk profile calculation, which was previously disconnected from the save path, is now wired into the assessment pipeline.

Custom risk thresholds set via the Set Thresholds dialog are now persisted across application restarts. Previously, thresholds reverted to defaults every time the application was reopened.

The cross-method comparison matrix now groups tool columns by domain with expand/collapse controls, defaulting to domain-average columns. Sidebar cards consolidated from six to three (Statistics, Trend Direction, Data Quality), and the column ratio adjusted from 2:1 to 3:2 for better readability. Assessment area and assessor are now stored as direct fields, replacing fragile title-based string matching.

The Project Management Overview tab now shows High-Risk Findings count from cardinal tools only instead of a misleading Average Risk gauge. Physical, Cognitive, and Reliability distribution gauges filter to validated cardinal tools only. Duplicate Risk Trends Over Time charts removed, leaving a single authoritative instance on the Risk Analytics page.

The GOMS editor page now uses theme-aware resource brushes instead of hardcoded white backgrounds and forced light theme, making it consistent with the rest of the application in both dark and light modes.

All content dialogs across the application now have correctly styled primary action buttons - affirmative actions in blue, destructive actions in red - ensuring a consistent and polished dialog experience throughout.

The CWA Decision Ladder has been corrected to the canonical 16-node Rasmussen/Vicente template with properly named node types matching published academic sources. Node positioning, centre topology, and structural connecting lines with directional arrowheads have all been corrected to show the proper flow structure.

Fixed the ATWIT tool saving risk scores on a 0-10 scale instead of the expected 0-100 scale, which caused cognitive workload scores to appear approximately ten times lower than their true value on the dashboard.

Fixed RULA grand score normalisation where a score of 2 (acceptable, no action required) was incorrectly displaying as 25/100 due to a hardcoded switch table overwriting the correct linear formula.

Fixed the same double-save pattern in REBA where the editor page overwrote the project service's correct normalisation formula with a hardcoded mapping, producing misleading risk values on the dashboard.

Corrected the Snook Tables normalisation formula to correctly map percentage-acceptable values to the 0-100 risk scale using an inverse relationship, and removed the conflicting 3-band mapping from the editor page.

Removed redundant duplicate save calls from the THERP, SPAR-H, and SHERPA editor pages that risked producing divergent values from the authoritative project service calculations.

Fixed the 3D workspace view displaying items mirrored horizontally. The camera-behind view in a right-handed coordinate system was causing workspace positions to appear reversed.

Fixed the elevation angle formula using only forward distance instead of true horizontal distance. Fixed the calculator using the box corner instead of the box centre for angle and distance calculations. Custom equipment placement now correctly uses the undo/redo command system.

Fixed the delete button executing a double removal. Locked items can no longer be accidentally deleted. Added protection against rapidly opening multiple dialog instances. Canvas background colours and dialog themes now correctly match the application theme in both light and dark modes. RAG indicators use theme-aware brushes.

Removed invented cross-domain multipliers from the risk heatmap that fabricated scores in domains where no tool data existed. Removed the hardcoded quality indicators panel (Consistency, Completeness, Data Quality, Review Compliance) that displayed fabricated values not derived from actual project data. Fixed tool category inconsistencies in the heatmap.

Fixed the Project Summary health status using different hardcoded thresholds (40/70) than the Risk Analytics page (30/70), causing the same score to be classified differently in different views.

Fixed editable session name fields showing invisible white text on a white background when hovered in dark mode. The hover state now correctly applies a grey background using theme-aware brushes.

Replaced all em dash characters across the application with standard hyphens for consistent typography. Standardised the update service URL to use the canonical non-www form matching the licence validation endpoint.

Audited ErgoTrace, ErgoStrata, ErgoLoop, ErgoGaze, and ErgoNeuro for the double-save pattern and removed redundant duplicate saves, ensuring each tool has a single authoritative save path through the project service.

A 40-point audit of the ErgoDesign workspace editor verified coordinate conventions, angle formulas, threshold values, serialisation completeness, and crash resilience, resulting in 11 defect fixes.

A comprehensive audit of the risk scoring system across all 35 assessment tools identified that only 9 of 16 scoring tools had defensible cardinal formulas. This audit drove the full Risk Analytics rebuild and normalisation corrections in this release.

Added thread safety protections to the mutable list properties on AppState (Assessments, AssessmentLinks, Records) that are accessed by both background services and the UI thread.

Addressed 15+ empty catch blocks across licence, persistence, and network code by adding proper error logging via ErrorHandlingService, ensuring failures are captured rather than silently discarded.