# Reality Engine tests Tests are defined by **markdown files** with a single fenced code block (language `test`) containing JSON. No hardcoded test list: add a new `.md` file to add a new test. - **TEMPLATE.md** — Schema and param reference; not run. - **run_tests.py** (in parent directory) — Discovers `*.md` here, parses the `test` block, runs the script with `params`, writes results to `results/<test_name>/`. Use `--progress N` to print step progress every N steps for long runs. - **rst_reality_engine_test_suite.py** (in parent directory) — Stress tests: MOND radial scan and 2D lensing. `--run all --micro` for a minimal run (~1–2 min); `--run all` (no flags) for the **long run**: MOND 200 nodes / 300 steps, log_radial 80/250, lensing 35×35 / 120 steps, EFE 80 steps (8 ω for gradient), flicker 25×25 / 80 steps. `--progress N` for step logging. Results and summary in `results/`; see `results/Micro run interpretation.md` for interpretation of a micro run. - **run_rst_proofs.py** (in parent directory) — Validation script: runs **seven** RST proofs (Litmus 3.0, JWST collapse, True EFE, Flicker, **External Mass Node**, **Coarse-grain**, **Weak format**); prints PASS/FAIL. Use `--micro` for a fast run; `--a0-only` to verify a0 calibration (no hardcoded 1.2e-10). - **Test suite runs:** `--run mond_scan | lensing | efe | log_radial | litmus | suppressor_efe | flicker | collapse | collapse_2 | global_noise_efe | efe_n | efe_omega | efe_omega_gradient | repulsion | external_mass_node | coarse_grain | weak_format | all`. New: **external_mass_node** (EFE from node `broadcasts_omega`), **coarse_grain** (`coarse_grain_super_node`), **weak_format** (Four-Force weak transition). - **Litmus 1.0, 2.0, 3.0** — All three variants are implemented. Litmus 1.0 and 2.0 are expected to FAIL (scale invariance); Litmus 3.0 (resistance matching) is expected to PASS. Run all: `rst_reality_engine_test_suite.py --run litmus --litmus-version all --micro`. - **Boundary tests** — Systematic stress tests with large graphs and long runs (~30–60 min each). **[[Boundary tests (systematic, ~1hr each)]]** describes objectives and configs. Run from parent directory: `python run_boundary_tests.py --test 1` (single test) or `python run_boundary_tests.py --test all --progress 25`. See TEMPLATE.md for the JSON format and how to run. Test .md files can include an **Implementation** section with the Python entrypoint and a short code snippet so what each test does is transparent (see efe.md, log_radial_mond.md, solar_system.md).