refactor(status): move benchmark status compaction into bounded read models

[huangruiteng]

Change

Q4 benchmark-sensitive completion slice: the benchmark status compaction block moves out of loopx/status.py into two bounded benchmark read-model modules:

• loopx/benchmarks/read_models/benchmark_status_compaction.py
• loopx/benchmarks/read_models/benchmark_status_runner.py

loopx.status now keeps thin compatibility wrappers for compact_benchmark_run, compact_benchmark_post_launch_materialization, and the private post-launch materialization helper. loopx/status.py drops from 2924 lines to 1392 lines, satisfying the M6 status.py < 2000 gate.

The build_skillsbench_post_run_debug_gate reexport is preserved as an audited compatibility entry with explicit test evidence.

Surfaces

• loopx/benchmarks/read_models/benchmark_status_compaction.py
• loopx/benchmarks/read_models/benchmark_status_runner.py
• loopx/status.py
• tests/benchmarks/read_models/test_skillsbench_post_run_debug.py

Validation

• Full pytest: 2306 passed, 2 skipped.
• Focused benchmark/status/import-boundary/maintainability: 45 passed.
• Ruff: All checks passed.
• loopx canary premerge --from-git-diff: selected 17 checks passed; one benchmark_sensitive manual hold is resolved by this owner review.
• Two benchmark-run-ledger smoke tests fail identically on clean origin/main; they are pre-existing and unrelated to this PR.

No new failures or skips beyond the pre-existing suite. No manual holds after owner review.

Routing

Route: codex-side-bypass; not assigned to codex-quality-qualification.

[huangruiteng]

Review: APPROVED

head ad3f0686a。这是对 PR #2978 的完整 owner review。该 PR 是 M6 Q4 的 benchmark-sensitive 收尾片：benchmark status compaction 从 loopx/status.py 迁入两个 bounded benchmark read-model 模块，loopx/status.py 降到 1392 行并保留薄兼容壳与 audited reexport。

验证

• 全量 pytest：2306 passed, 2 skipped
• 聚焦 benchmark/status/import-boundary/maintainability：45 passed
• ruff：通过
• loopx canary premerge --from-git-diff：selected 17 checks passed；唯一 benchmark_sensitive manual hold 由本 owner review 覆盖
• 两个 benchmark-run-ledger smoke 在干净 origin/main 上也失败，与本 PR 无关
• GitHub checks：dependency-review / pytest / build 全部通过

结论

无 blocker。benchmark status compaction 现在位于 bounded benchmark read-models，公共 facade 行为保持。按仓库 merge policy 路由合并。