Add comprehensive test suites for answer key validation, API functionality, and code linting

- Introduced `test_answer_key.py` to validate the correctness of answer keys against expected responses, ensuring both false and true answers are graded appropriately.
- Created `test_answer_values.py` to verify the accuracy of answer keys by recomputing values and comparing them against the provided `answers.json`.
- Implemented `test_api.py` to test the HTTP API for suite management, ensuring read-only enforcement and proper handling of custom suites.
- Added `test_code_lint.py` to perform static code analysis, ensuring that all linting rules are enforced without false positives on correct answers.
- Developed `test_collision.py` to guard against prompt mismatches due to filename collisions across suites.
- Created `test_gguf.py` to validate GGUF header parsing and model classification, including handling of malformed input.
- Introduced `test_linter.py` to ensure report rendering preserves content integrity and correctly fences code.
- Added `test_reporting.py` to verify report naming conventions and the accuracy of per-suite score breakdowns.

README.md

@@ -300,6 +300,18 @@ Reports are written to `results/automated_report_<model>.md`.
 
 ---
 
+## Tests
+
+```bash
+python tests_py/run_all.py     # 195 assertions, no dependencies
+cd web && npm test             # 29 UI assertions
+```
+
+`tests_py/test_api.py` needs a running server and skips cleanly without one.
+Details in [`tests_py/README.md`](tests_py/README.md).
+
+---
+
 ## Project structure
 
 ```
@@ -320,6 +332,7 @@ server/                 FastAPI backend wrapping the engine
 plugins/                Drop-in plugins — _skeleton.py is the starter template
 plugin_api.py           Stable types plugins import
 plugin_system.py        Plugin discovery and hook dispatch
+tests_py/               Python tests — python tests_py/run_all.py
 web/                    React + TypeScript interface (Vite)
 models/                 Drop .gguf or MLX weights here (auto-discovered)
 .core/suites/<slug>/    Built-in suites, read-only