SupraTix GmbH – Releases

1.0.15.8

UPSKILL MANAGER – CHANGELOG
Period: 24 June 2025 – 18 September 2026
Scope: mint/upskill_manager (~700 commits incl. dependency updates)

========================================
HIGHLIGHTS
========================================
- New Upskill Manager dashboard ("Talent Intelligence Workspace") with live KPIs, tool tiles and a skill-bridge preview.
- Central Skill Catalog and interactive Skill Graph (D3), including an ontology/knowledge-graph view and VectorGraphRAG-based behavior-anchor provenance.
- AI features: AutoTag skill suggestions for content, SkillProbe micro-assessments, Skill Passport, and an explainable upskilling-recommendation engine.
- Measurements rebuilt: self vs. external assessment (Selbstbild/Fremdbild) with radar view, invitations by mail, measurement methods generated from competence models, and a new processing browser.
- Target / job profiles: assign target profiles to people (no duplicates), view the most recent target profile, and file profiles into folders.
- Strategizer & position network: React Flow graph of job profiles, strategic-alignment overview, critical roles, succession and talent-acquisition insights.
- Formalizer: behavior anchors and learning objectives can be allocated directly, and you can see where each learning objective is used.
- Full German localisation of the app's visible text, plus performance and tenant-security hardening throughout.

========================================
NEW FEATURES
========================================

DASHBOARD & NAVIGATION
- New dashboard with KPIs (active competence models, active job profiles, completed measurements), tool tiles with short descriptions and an empty-state hint for new workspaces (Jul–Aug 2026).
- New base_v3 layout and dashboard stylesheet. Navigation in the competence model, formalizer, measurement and strategizer menus reworked, with a new Skill Catalog sub-menu (Aug 2026).
- Continuous-learning insights on the dashboard, scoped per tenant, with an empty-state prompt (Sep–Oct 2025).
- Clear "no valid licence" state that links to the store.

COMPETENCE MODELS (KOMPETENZMODELLE)
- The gallery has server-side search and filters. You can now install a model from the gallery, with a licence check and a store hand-off (Jan & Apr 2026).
- Model editor: competence descriptions save automatically, new competences show up right away, behavior anchors can be added inline and are linked to competence items (Jul 2025).
- Recommendations for extending a competence model, generated by AutoTag (Mar 2026).
- Create a competence-model product directly from the workflow (Aug 2026).

FORMALIZER (CONTENT -> ACTION-ORIENTED SKILLS)
- Add behavior-anchor-based learning objectives directly (Apr 2026).
- Shared competence-reference layer. Anchors and learning objectives are resolved the same way in the Formalizer and in Factory metadata, so both stay in sync after store-connect releases (Sep 2026).
- Objective usage view on all four formalizer steps shows where each learning objective is currently assigned (Sep 2026).
- Course list shows formalizer progress, search by title or course ID, and clearer empty states. Step 3 now uses shared logic for level and tolerance (Sep 2026).
- Evaluation-question workflow hardened: CSRF handling, recovery when the session expires, saving correct answers (Dec 2025).

TARGET & JOB PROFILES (IST-/ZIEL-STELLENPROFILE)
- Target-profile checkboxes and accordion states are kept after a reload. Anchors are pre-selected from a user's competence objectives (Jun 2025).
- Assign a target profile to a person with a target completion date. Duplicate assignments are now prevented and existing duplicates were removed (Aug 2026).
- New "most recent target profile" card for each person (Aug 2026).
- Target Competency Profiles REST API (list, detail, mutate). Validation of scale, level and weight was tightened, and a limit on the number of objectives was added (Apr & Sep 2026).
- Job-profile folders: create, rename, dissolve and move profiles between folders. No database migration is needed (Sep 2026).

MEASUREMENTS (MESSMETHODEN & MESSUNGEN)
- Measure user behavior with questionnaires, plus transcript-based competence extraction and evaluation that runs as a background job (Jun–Jul 2025).
- Setup workflow for measurement methods (4 steps). Measurement methods can be generated automatically from competence models, one question per behavior anchor (Oct–Nov 2025).
- Measurement gallery: search, vendor filter, real statistics on the cards, booking and checkout straight from the gallery (Nov 2025 – Jan 2026).
- Self vs. external assessment (Selbstbild/Fremdbild): comparison table, radar chart and difference classes. Invitation mails in HTML and plain text with an anchor and scale summary, plus timeline entries (Jun 2026).
- Employee selection now has search (name, e-mail, position). Only licensed measurement methods are offered (Jun 2026).
- The measurement scale now comes from the competence model (max value, step count) (Aug 2026).
- New processing browser with modals to create a measurement method, load anchors dynamically and set an optional anchor value (Sep 2026).
- Results are aligned with job profiles and show questions that have no behavior anchor (Oct 2025, Aug 2026).

PATHWAYS & POSITION NETWORK (POSITIONSNETZWERK)
- Learning-pathway generator, pathway API endpoints, goal selection and overview pages, with permission checks (Aug 2025).
- React-based profile editor and skill-gap dashboard. Positions open in a modal, and pathway progress is shown per user (Oct 2025, Aug 2026).
- Layouts of the competence network are saved per profile (Feb 2026).

STRATEGIZER
- React Flow visualisation of job profiles, a strategic-priority toggle and highlighted critical roles (Oct 2025).
- Career-path, succession-planning and talent-acquisition insights. Graph summary metrics such as average path length (Oct 2025).
- Strategic alignment overview with inline edit links (Oct 2025).
- Skill-gap calculation takes strategy into account (strategy-to-skill overlay), and you can simulate what a strategy change would do (Feb–Mar 2026).

SKILL CATALOG & SKILL GRAPH
- Central Skill Catalog: create and maintain skills, set parent groups, show or hide inactive skills, see linked models, filter by competence model. Includes a REST API (Apr 2026).
- Interactive Skill Graph (D3): node details, legend, rebuild option, lasso selection (Alt + drag) and path finding (Shift + click) (Apr–Jul 2026).
- Ontology / knowledge-graph concept section. Tenant-specific graph product with allocated product contexts (Jul 2026).
- Precomputed skill-graph dataset, embeddings, and a live skill graph API that updates through signals (Feb–May 2026).
- VectorGraphRAG embeddings on behavior anchors. The graph is projected through the HR-AI source projection (Sep 2026).

AI & RECOMMENDATIONS
- AutoTag: AI skill suggestions for content items, with reasoning and a suggested level. You can approve or reject them, process the backlog, generate evaluation tasks, and have learning objectives assigned automatically (Mar & May 2026).
- SkillProbe: adaptive micro-assessments, hardened with voice evidence from timeline calls (Mar 2026).
- My Skill Passport: traceable skills with evidence, confidence and a share link (Mar 2026).
- Upskilling recommendations API: explainable bundles covering skill gaps, learning resources, inventory evidence and explain badges (May 2026).
- Job recommendations show salary comparison, skill-match details and default suggestions for users without a profile. A skill-assessment quiz was added (Aug 2025).
- Cosine-similarity matching with visualisation. Weighted competence and fit-score calculation (Jun 2025).

EXPORTS & APIS
- Knowledge Graph API (CRUD and queries) moved into the Upskill Manager (Jan 2026).
- Semantic export as JSON-LD and RDF/Turtle, with a new export page and format picker (Nov 2025, Jul–Aug 2026).
- Behavior-anchor API per competence model (Jun 2026).

========================================
IMPROVEMENTS
========================================
- Performance: fewer and batched queries for anchor values, fit score, cosine similarity, position availability, job-profile recommendations, questionnaire cloning and transcript processing. Faster request paths for pathways and public pages (Sep 2025, Aug 2026).
- Localisation: all visible template text wrapped for translation and translated into German. UI wording made consistent (May–Aug 2026).
- UI: measurement-method details restyled. Semantic-export view moved to Fomantic UI. Status element after saves. Accessibility improved (2025–2026).
- Code quality: docstrings added throughout, timezone-aware dates, template-tag loads cleaned up, shared helpers reused (May 2026).

========================================
SECURITY & PERMISSIONS
========================================
- All relevant views now require a module subscription (Dec 2025).
- Permission classes for admin and catalog APIs. Non-admins only see their own records in the upskill API (Feb–Apr, Aug 2026).
- Tenant delivery policy enforced, and dashboard stats scoped per tenant (Oct 2025, Aug 2026).
- CSRF tokens added to AJAX calls (course step 3, formalizer questions, competence editing).
- jQuery updated from 3.6.0 to 3.6.4 (Sep 2026).

========================================
BUG FIXES (SELECTION)
========================================
- Deletion loops, recursive checkbox flags and anchor-value checks when an anchor is missing (Jun 2025).
- Rendering and markup errors in competence-model step 2. Saving of behavior anchors and descriptions (Jul 2025).
- Username shown on the IST profile. Fallback lookup for job profiles. Error in strategizer labels (Jul & Oct 2025).
- Measurement methods: missing IDs, vendor lookups, question count, template errors, invalid measurement method ID (Nov 2025 – Jun 2026).
- Default tolerance display. Assigning learning objectives to relations. Missing subscription when a job profile is deleted (Nov–Dec 2025).
- Target-profile assignment, OPC UA refresh, and database and frontend regressions (Aug 2026).

========================================
MAINTENANCE
========================================
- Tests moved into the central test suite, with an isolated tenant-database suite (Jun & Aug 2026).
- Front-end dependencies kept current: React 19.2, react-router 7, Vite 8.3, ESLint 10, plus security updates (js-yaml, minimatch, picomatch, brace-expansion, flatted, ajv, rollup).

## What's Changed

* feat(movement): define signed packages and canonical pinned process contracts by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/399

* feat(commissioning): persist tenant-bound first-login wizard progress by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/398

* fix(lvgl): admit commissioning wizard in the actual page router by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/400

* feat(commissioning): synchronize immutable tenant device dossiers by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/401

* fix(installer): refuse host replacement of journal-managed releases by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/402

* fix(process): enforce one Petri execution and recovery owner across runtimes by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/405

* feat(vision): wire commissioned full-tray localization into Petri states by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/403

* fix(lvgl): show feedback for rejected tenant keypad input by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/404

* fix(motion): close unsigned motor-enable entry points by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/412

* fix(ota): pin publisher trust through final release promotion by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/413

* fix(xarm): exclude concurrent production SDK owners by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/407

* feat(boundary): install reviewed dual-camera validation from wizard by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/409

* fix(ota): scope archive credentials and verify remote release tags by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/411

* fix(commissioning): reject FIFO evidence without blocking startup by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/414

* feat(ota): preserve acknowledged journals during host maintenance by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/416

* feat(petri): deliver frozen tenant evidence with exact durable receipts by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/431

* feat(petri-example): compose the editable full-tray process from pinned modules by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/430

* test(ui): isolate legacy cancellation import stubs by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/429

* feat(process-editor): edit typed tenant drafts and submit immutable definitions by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/428

* fix(commissioning): post dossiers to the mounted Device Manager route by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/427

* feat(petri): persist verified originals and closure-bound run manifests by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/425

* feat(petri): bind reviewed lid cycles to canonical ownership and closure proof by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/423

* fix(recovery): wire signed reviewed Home with measured arrival by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/426

* feat(ota): revalidate privileged admission across release switches by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/424

* test(teach): verify abrupt writer death preserves audit-only proposals by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/422

* docs(nvidia): distinguish runtime support from deployment evidence by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/421

* fix(ota): preserve authoritative terminal lifecycle reports by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/418

* feat(commissioning): draft reviewed one-use manual-teach authority by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/417

* feat(gallery): install tenant-bound verified movement versions from LVGL by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/419

* fix(screen-share): report click rejection reason, verify XTest delivery by @tobiasgoecke in https://github.com/Supratix/labcobot/pull/432

Full Changelog: https://github.com/Supratix/labcobot/compare/1.0.15.6...1.0.15.7

Signed AINA pilot OTA containing the exact proven 1.0.3 application runtime, repackaged as 1.0.7 for Supratix-only delivery. No robot-control or physical Teach Mode changes. Source revision: 84f66827d4b8e39c9ecdc10675b5a3b6bb67155e. Artifact SHA-256: 108b56397bfecd42bd9d7d3240f7651aaf2c36b5059a731c5ceb894156ce1158.

LabCobot OTA 1.0.2. Fail-closed handling for malformed or non-iterable updater phase metadata and malformed admission lease durations. Automatic OTA remains non-authorizing and requires a stationary controller, fresh cameras, no physical owner, and generation-bound admission evidence. Source revision: 035fe998db3c2e0aebcc8769ac0851237f9937db. Artifact SHA-256: 53cc9484377d9cfb92732e796ea0f7a6c921c2e80a1a1e7652879c6055692f23.