L'investigation initiale supposait des customFields JSON dans les
skeleton_*_requirements ; en realite SkeletonStructureService traduit
les customFields du payload ModelType en entites CustomField stockees
dans la table custom_fields. Le SQL est donc un simple SELECT DISTINCT.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Spec validee : endpoint backend qui agrege les noms existants (table
custom_fields + JSON dans skeleton requirements), composant
CustomFieldNameInput qui wrap SearchSelect en mode creatable, cache
module-level partage entre toutes les instances, invalidation apres save.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Backend:
- Enrich machine snapshot with componentLinks/pieceLinks/productLinks
- Detect link add/remove in MachineAuditSubscriber onFlush
- Add link diff comparison in restore preview
- Add link restoration in applyRestore for machines
- Add integrity warnings for missing linked entities
Frontend (submodule update):
- Single save button replacing auto-save-on-blur
- Link versioning display in version list and restore modal
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- 3 MCP resources: schema, roles, stats
- docs/mcp/README.md with full user guide (config, tools catalogue, workflows)
- .mcp.json for Claude Code stdio transport
- Design spec and implementation plan
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Address review findings: drop Groups attribute, add clone logic,
specify PATCH payload format, list frontend functions to update,
add validation and test cases.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
