What problem does it solve? Modeling prefilters in a Thinkwise Software Factory requires knowing the exact entity structure (tab_prefilter, tab_prefilter_grp, tab_variant_prefilter_overview), the difference between column-based and query-based filter logic, and subtle behaviors like match-all vs. match-any semantics and per-context prefilter states. This Skill provides a verified reference so an AI agent working through an MCP connector creates, inspects, and troubleshoots prefilters correctly instead of guessing entity names or field values. ## Core Features & Use Cases - Verified entity reference: Documents the fields, enums, and write behavior of tab_prefilter, tab_prefilter_grp, and tab_variant_prefilter_overview, confirmed against a live Software Factory model. - Column-based vs. query-based guidance: Explains when to use a prefilter column (validated, no SQL) versus a hand-written SQL fragment aliased as t1, including the verified gap where prefilter column rows are not exposed as an entity set. - Group semantics and states: Covers match-all vs. match-any combination logic, mandatory and exclusive groups, the five prefilter states (including On hidden for data authorization), and variant override rules. - Use Case: A user asks the AI to add an "Overdue" prefilter to an orders table. The Skill directs the agent to plan the prefilter for confirmation, write the query fragment t1.due_date < getdate(), set an icon and tooltip, place it in a sub_menu group, and verify translations across all branch languages. ## Quick Start Ask the AI to create a prefilter on a specific table in your Software Factory model, for example to show only overdue records, and let it plan and stage the prefilter through the MCP connector.