fix(records): repair unreachable DONE owner commits

.agents/coordination.md

@@ -118,6 +118,23 @@ without the selected contention proof for their entire run are discarded.
 
 ## Active claims
 
+**Canonical DONE-owner reachability repair (`KV-PREFIX-CACHE`,
+`SAMPLE-LOGPROBS`, `SPEC-DFLASH`,
+`MODEL-SPEC-qwen3-dflash-dflash-qwen3-for-causal-lm`,
+`KERNEL-ATTN-DFLASH-BLOCK`, `KERNEL-ATTN-DFLASH-PAGED-BLOCK`, 2026-07-31,
+`CLAIM-RECORD-DONE-OWNER-REPAIR`).** Codex (GPT-5), isolated worktree
+`/home/mudler/.cache/sdd/localai-org-maint-bot-vllm.cpp/codex-cpu-regression-fix`,
+branch `codex/cpu-regression-fix`, base `upstream/main` `0d014016`. CPU-only
+governance repair: `scripts/check-agent-record.py` currently rejects six `DONE`
+rows because their three owner SHAs are not Git objects; replace them with the
+reachable commits that actually introduced the binding `DONE` closures. Owns
+only those six Owner cells, this claim, the minimal append-only state entry,
+and the required no-lifecycle-change checkpoint notes in `docs/STATUS.md` and
+`docs/BENCHMARKS.md`. No source, test, build, lifecycle, benchmark, README,
+roadmap, ledger, or support-claim change; verification is the agent-record
+checker and its mutation suite plus the remaining record/document checkers.
+No GPU, model, or compiler is required.
+
 **Stable C ABI contract spike (`SERVE-C-ABI`, 2026-07-31,
 `CLAIM-SERVE-C-ABI-SPIKE`).** Codex (GPT-5), isolated worktree
 `.worktrees/serve-c-abi-spike`, branch `codex/serve-c-abi-spike`, base

.agents/state.md

@@ -34430,3 +34430,5 @@ PARTIAL on two further unrelated test-only GCC 12 `-Wrestrict` diagnostics in
 `test_deepseek_v2_paged_engine.cpp` (actively MLA-owned) and
 `test_glm4_moe_lite_paged_engine.cpp`; those files were not touched, so full
 CTest cannot run from this incomplete build. No GPU, model, or benchmark ran.
+
+**Canonical DONE-owner reachability repaired (2026-07-31, `CLAIM-RECORD-DONE-OWNER-REPAIR`).** Records-only correction: the six checker-reported `DONE` owner cells now name the reachable commits that introduced their binding closures (`KV-PREFIX-CACHE` `a41af480`, `SAMPLE-LOGPROBS` `ecda3ce1`, and the four DFlash rows `489a7544`). Lifecycle states, evidence, support claims, source, tests, and benchmark results are unchanged.

docs/BENCHMARKS.md

@@ -10967,3 +10967,10 @@ diagnostics outside this diff. The row remains
 `ANCHOR-BACKFILL`; W1-W5 in
 [the spec](../.agents/specs/c-api-library.md) name the remaining compatibility,
 no-throw, lifetime, and standalone-consumer gates.
+
+## Canonical record-owner repair (2026-07-31) - NOT APPLICABLE
+
+Six existing `DONE` matrix rows now point to reachable commits that introduced
+their binding closure states. This records-only metadata repair changes no
+lifecycle state, implementation, workload, or accepted benchmark result, so no
+new performance number is applicable or claimed.

docs/STATUS.md

@@ -2059,3 +2059,5 @@ _(Laguna W11 go/no-go: decode is GPU-COMPUTE-bound (GPU-busy 2.56s ≈ sync 2.59
 **qwen3_5 A3 W3b LANDED (2026-07-31) — A3 COMPLETE.** Per-expert keep-quant MoE loop → 3 grouped vt::MatmulBTQuantGrouped launches (VT_QWEN35_GROUPED_MOE default-ON). BYTE-EXACT: DGX APEX-Compact grouped(=1) vs per-expert(=0) continuations byte-identical + strict-passes golden. Shared descriptor with Laguna W9. See spec §W3b.
 
 **Fold-plan A3 tier CLOSED (2026-07-31).** Both keep-quant models (Laguna W9 + qwen3_5 GGUF W2/W3a/W3b) route their routed-expert MoE through the shared vt::MatmulBTQuantGrouped descriptor, byte-exact gated. See .agents/specs/arch-fusion-fold-plan-2026-07-30.md A3.
+
+**Canonical record-owner repair (2026-07-31).** No lifecycle or support status changed. Six existing `DONE` rows now point to their reachable binding closure commits; this is records-only metadata maintenance.