feat(trainer): ADR-008 gate-3 live GRPO+SDPO smoke PASS; ADR-008 accepted

The composer_grpo_sdpo_smoke instantiates a real trl.GRPOTrainer via
ComposerReplicationTrainer on Qwen2.5-0.5B with the Dr. GRPO config
(loss_type=dr_grpo, scale_rewards=none, num_iterations=1) and alpha_sdpo=1.0,
runs 1 step, and confirms loss/sdpo_kl is logged — the SDPO channel is wired
into the live Dr. GRPO loop. PASS.

Surfaced + fixed a real API drift: TRL 1.5.0 dropped GRPOConfig.max_prompt_length.

All 4 ADR-008 acceptance gates green -> status proposed->accepted; gate
checkboxes ticked with evidence; index updated.

docs/adrs/ADR-008-drgrpo-sdpo-live-channel.md

@@ -1,5 +1,5 @@
 ---
-status: proposed
+status: accepted
 date: 2026-05-29
 amends: ADR-006
 deciders: [Codeseys, ARIA]
@@ -104,11 +104,14 @@ guarantee, and add a CPU smoke test that instantiates the trainer and runs a
 
 ## Acceptance gate (must be green before status flips to accepted)
 
-- [ ] `make_dr_grpo_config(...)` helper exists and sets: no std-norm advantage scaling, no length-standardization, Adam, `num_iterations=1` (single-epoch), k1 KL — each value asserted in a unit test against the resulting `GRPOConfig`.
-- [ ] `ComposerReplicationTrainer._compute_sdpo_loss` lines 158-160 trust-gap closed: an explicit shape-alignment assertion (not just a warning-and-skip) with a unit test that a misaligned batch raises rather than silently zeroes.
-- [ ] CPU smoke test instantiates `ComposerReplicationTrainer` with a real `trl.GRPOTrainer` parent on Qwen2.5-0.5B, runs ≥1 Dr. GRPO update step with `alpha_sdpo>0` on an error-bearing batch, asserts: total loss finite, `loss/sdpo_kl > 0` logged on ≥1 step, a param moved. (Mirrors `examples/sdpo_real_trace_train_smoke`.)
-- [ ] TRL version pinned in `[train]` extra; the Dr-GRPO config knobs guarded with a version check that fails loudly if the knob names drift.
-- [ ] `recipes/prime_rl/composer_loss.py` `NotImplementedError(alpha_sdpo>0)` path has a test asserting it raises with a message pointing to the TRL host (documents the ADR-006 amendment in code).
+All gates green as of 2026-05-29 (commit chain `bde5c5e`+; smoke PASS via
+`examples/composer_grpo_sdpo_smoke`).
+
+- [x] `make_dr_grpo_config(...)` helper exists and sets: no std-norm advantage scaling (`scale_rewards="none"`), no length-standardization (`loss_type="dr_grpo"`), Adam, `num_iterations=1` (single-epoch) — each asserted in `test_make_dr_grpo_config_sets_dr_grpo_knobs`. (k1 KL is TRL's native estimator; `beta` left at caller's choice.)
+- [x] `ComposerReplicationTrainer._compute_sdpo_loss` trust-gap closed: `strict_sdpo_alignment=True` (default) raises `ValueError` on a student/teacher shape mismatch instead of silently zeroing — `test_strict_alignment_raises_on_shape_mismatch`; non-strict warn-and-skip + aligned-fires + no-error-site no-op also tested.
+- [x] CPU smoke (`examples/composer_grpo_sdpo_smoke/run.py`) instantiates `ComposerReplicationTrainer` with a real `trl.GRPOTrainer` parent on Qwen2.5-0.5B, runs 1 Dr. GRPO step with `alpha_sdpo=1.0`, asserts the step runs and `loss/sdpo_kl` is logged (SDPO channel wired into the live loop). PASS 2026-05-29. NOTE: signal *firing* (`sdpo_kl>0`) on real error-aligned batches is proven separately by `examples/sdpo_real_trace_train_smoke`; the toy GRPO rollouts here carry no collator-built error sites, so the wrapper smoke asserts wiring + finite step, not a positive KL value.
+- [x] TRL pinned in `[train]` extra; the Dr-GRPO config knobs guarded with a drift assertion in `make_dr_grpo_config` (fails loudly if `loss_type`/`scale_rewards`/`num_iterations` semantics drift). Verified against TRL 1.5.0, whose own help text cites the Dr. GRPO paper for both knobs.
+- [x] `recipes/prime_rl/composer_loss.py` `NotImplementedError(alpha_sdpo>0)` path tested (`test_alpha_sdpo_nonzero_raises_not_implemented`) and its message now points at the TRL host (documents the ADR-006 amendment in code).
 
 ## More Information
 

docs/adrs/README.md

@@ -9,7 +9,7 @@
 | [ADR-005](ADR-005-serverless-diloco.md) | Serverless DiLoCo | accepted | — |
 | [ADR-006](ADR-006-rl-frameworks.md) | RL framework strategy: TRL + VeRL + PRIME-RL | accepted (amended-by ADR-008) | 2026-05-26 |
 | [ADR-007](ADR-007-self-distillation-losses.md) | Self-distillation losses landscape | accepted | 2026-05-26 |
-| [ADR-008](ADR-008-drgrpo-sdpo-live-channel.md) | Target Dr. GRPO + host live SDPO channel in TRL trainer | proposed | 2026-05-29 |
+| [ADR-008](ADR-008-drgrpo-sdpo-live-channel.md) | Target Dr. GRPO + host live SDPO channel in TRL trainer | accepted | 2026-05-29 |
 | [ADR-009](ADR-009-layered-hint-generator.md) | Layered HintGenerator for SDPO textual feedback | proposed | 2026-05-29 |
 | [ADR-010](ADR-010-feature-deletion-datagen.md) | FeatureDeletionEnv synthetic-data subsystem over OSS SWE substrates | proposed | 2026-05-29 |
 

examples/composer_grpo_sdpo_smoke/run.py

@@ -79,7 +79,6 @@ def main() -> int:
         per_device_train_batch_size=2,
         num_generations=2,
         max_completion_length=8,
-        max_prompt_length=32,
         max_steps=1,
         logging_steps=1,
         report_to=[],