CI: unpinned ruff drifted to 0.16 — lint/format now fail repo-wide

Nobody has claimed this yet.

Every fresh PR currently fails the `run_tests` workflow at **Check styling** / **Linting**, independent of its content. Main last passed `Run tests` on 2026-07-16; since then unpinned `ruff` resolved to 0.16.x, which

- formats Python code fences inside markdown (flags `.claude/skills/add-ssm-model/SKILL.md`), and
- expands the effective default lint rules: `ruff check .` now reports ~300 findings concentrated in pre-existing files (`tests/rl/test_rl_config.py` 25, `tests/rl/test_assembled_model.py` 24, `tests/rl/test_rl_simulator.py` 20, `ssms/support_utils/utils.py` ~29, …; top rules I001 ×92, RUF012 ×81, UP006 ×37, RUF100 ×33, UP035 ×19).

PR #316 pins `ruff>=0.15.1,<0.16` (both dev listings) so CI reflects PR content again — 0.15.22 passes format+lint cleanly. To migrate: run `ruff check . --fix` + `ruff format .` under 0.16, triage the non-autofixable rules (RUF012, BLE001, TRY004, …) or add them to the existing `ignore` list, then drop the pin. Alternatively pin an explicit `select` to freeze the rule set against future default changes.

Start with the run_tests workflow and the two dev dependency listings, reproducing the failure with Ruff 0.16 and comparing it with 0.15.22. Review the reported findings in tests/rl/test_rl_config.py, tests/rl/test_assembled_model.py, tests/rl/test_rl_simulator.py, and ssms/support_utils/utils.py; done means the chosen pin or rule configuration lets format and lint pass without unrelated failures.

Written by the indexing model from the issue text.