Display field, filter, sort, and validation tests are spread inconsistently across the test suite: **Current state:** - `field_selection_tests.rs` — display field tests for WorkItems, Issues, Epics, and MergeRequests in monolithic per-source tests - `project_tests.rs` — Projects source has its own file with filters, display fields, sorting, transforms, and validation all in one place (good pattern) - `merge_requests_tests.rs` — MR filter tests only, no display field tests (those are in `field_selection_tests.rs` and `graphql_tests.rs`) - `graphql_tests.rs` — mix of display field structural assertions, scope tests, and output format tests - Individual field files (`label_tests.rs`, `assignee_tests.rs`, `health_tests.rs`, etc.) — WorkItem filter tests split by field - `misc_tests.rs` — catch-all for subscribed, myReaction, and output format tests - `scope_tests.rs`, `semantic_tests.rs` — cross-cutting concerns **Problems:** 1. No consistent pattern: Projects has everything in one file, WorkItems splits across ~15 files, MRs split across 3 files 2. Display field tests for MRs and WorkItems live in different files than their filter tests 3. `graphql_tests.rs` mixes concerns (display fields, scope, output structure) 4. Hard to find all tests for a given source **Proposal:** Adopt the `project_tests.rs` pattern: one file per source containing all tests (filters, display fields, sorting, validation, transforms). Keep cross-cutting test files (`semantic_tests.rs`, `parser_tests.rs`, `scope_tests.rs`) for shared behavior.