[WS1][C3] Land forward config-invariance harness and backend provenance

.github/workflows/ci.yml

@@ -67,6 +67,7 @@ jobs:
         run: |
           python -m pytest rl_engine/tests/test_dispatch.py -v
           PYTEST_DISABLE_PLUGIN_AUTOLOAD=1 python -m pytest tests/test_attention_correctness.py -q -rs
+          python -m pytest tests/test_forward_invariance.py -q
 
       - name: Run Attention Ground-Truth Tests (CPU-safe)
         run: |

docs/design/ws1-c3-269-closeout-evidence.md

@@ -0,0 +1,64 @@
+# WS1 C3 (#269) closeout evidence
+
+**Parent:** #266 · **Depends on:** #267 / #268 · **Scope:** shared forward harness only
+
+## Acceptance map
+
+| #269 criterion | Evidence |
+| --- | --- |
+| Accuracy and invariance separate | `ForwardInvarianceReport.accuracy_reports` and `invariance_reports` |
+| Batch/chunk bitwise after logical unpadding | C1 `forward_invariance` resolver plus exact C2 logical-key validation |
+| C2 transforms | `build_config_matrix`: fixed 2×2 matrix, permutation, packing, left/right padding |
+| Diagnostics | tensor name, config pair, max/mean absolute error, max relative error |
+| Backend provenance | profile, requested/actual backend, candidate/kernel id, device, CC, dtype, seed, fallback reason |
+| Silent/cross-profile fallback | missing or mismatched provenance fails; CLI rejects candidate/profile mismatch |
+| Selected-logprob smoke | C1 `max_abs_dlogp`, `approx_kl0`, and `clipfrac0` verdict |
+| CUDA and Triton same schema | one API/CLI/report schema; both profile contracts are parametrically tested |
+| No private thresholds | all tensor and aggregate thresholds resolve through the C1 contract |
+
+CPU-safe contract regression:
+
+```bash
+python -m pytest -q \
+  tests/test_tolerance_contract.py \
+  tests/test_ws1_workload.py \
+  tests/test_forward_invariance.py \
+  tests/test_op_checks.py
+```
+
+Required-profile runtime examples (must run on CUDA hardware and must not be skipped):
+
+```bash
+python scripts/check_forward_invariance.py \
+  --op logp --candidate cuda \
+  --backend-profile cuda_bf16 --json
+
+python scripts/check_forward_invariance.py \
+  --op batch_invariant_logp --candidate triton \
+  --backend-profile triton_cuda_bf16 --json
+```
+
+The CLI exits red when CUDA is unavailable, a candidate is absent, the C2 node is
+`missing_required`, the compute capability cannot run a declared SM90 candidate, provenance
+does not match the profile, or any accuracy/invariance/logprob verdict fails.
+
+## Runtime verification
+
+Verified on NVIDIA GeForce RTX 3060 Laptop GPU (`sm86`) with PyTorch 2.8.0+cu128:
+
+| Gate | Result |
+| --- | --- |
+| Full pytest suite | `1524 passed, 121 skipped` |
+| Full pre-commit | trailing whitespace, EOF, YAML, large-file, black, isort, flake8 passed |
+| `cuda_bf16` / generic CUDA logp C3 matrix | passed; all invariance max-abs errors `0.0` |
+| `triton_cuda_bf16` / Triton batch-invariant-logp C3 matrix | passed; all invariance max-abs errors `0.0` |
+| CUDA operator accuracy check | passed, max absolute error `0.0287590` |
+| Triton operator accuracy check | passed, max absolute error `9.536743e-07` |
+
+The CUDA profile uses the manifest-declared generic CUDA logp candidate on SM86. No SM90
+candidate or fallback path is claimed on this device.
+
+## Parent boundary
+
+This closes only C3. It supplies the report and canonicalization contract that C10 must reuse.
+It does not claim the full-model, backward, KV-cache, or CI EXIT requirements of #266.

rl_engine/kernels/gtest/__init__.py

@@ -1,6 +1,16 @@
 # SPDX-License-Identifier: Apache-2.0
 # Copyright (c) 2026 RL-Kernel Contributors
 
+from .forward_invariance import (
+    AccuracyReport,
+    ConfigSpec,
+    ForwardInvarianceReport,
+    InvarianceReport,
+    LogprobSmokeResult,
+    TensorComparisonDetail,
+    assert_forward_batch_invariant,
+    build_config_matrix,
+)
 from .op_checks import CandidateSpec, OperatorCase, run_operator_suite
 from .tolerance import (
     BackendProvenance,
@@ -15,8 +25,16 @@
 )
 
 __all__ = [
+    "AccuracyReport",
     "CandidateSpec",
+    "ConfigSpec",
+    "ForwardInvarianceReport",
+    "InvarianceReport",
+    "LogprobSmokeResult",
     "OperatorCase",
+    "TensorComparisonDetail",
+    "assert_forward_batch_invariant",
+    "build_config_matrix",
     "run_operator_suite",
     "BackendProvenance",
     "ContractError",