…e current gateway builds (#641)

* Foundation catalog drift DD-1/DD-2: one Agents SDK disposition and the current gateway builds

- DD-1 (catalogs/landscape/foundation.json): the OpenAI Agents SDK gets one disposition, unqualified, with the missing
  layer entries; the managed API is distinguished from the SDK's configurable endpoint and native Codex wrapper.
  OmniRoute request compatibility stays unmeasured and AD-5 (codex_tool) stays on hold. The 2026-09-26 sweep returns
  and their missing-vote provenance are unchanged.
- DD-2 (catalogs/foundation/manifest.json): the gateway description names the builds on 20128 and 20129, their shared
  2f42a9ac base, and the clean 3.8.51 package's xhigh cap for gpt-6.1-sol versus upstream #15167's max, as dated
  observations (PR #637's decision).
- catalogs/us-equities/decision-index.json regenerated by scripts/catalog_decisions.py --write (a foundation file under
  a trading path, docs/lanes.md); two anti-pattern rows in docs/harness-defaults.md; decision record
  docs/decisions/2026-10-03-foundation-catalog-drift.md.
From the 0c Codex-practice audit (2026-10-02) items DD-1 and DD-2, cross-checked by GPT-6.1 Sol.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Register the catalog drift files (hot-file protocol, last commit)

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Keep the frozen native-clients verdict fields; reconcile the agent-sdks gate; register the layer-16 supplement

Hosted validate and verdict-review-gate refused 226f079: the native-clients alternatives[2] text, overturn arms and
verdict_overturn_when are fields of the frozen wave-20260922 layer-verdict row, which changes only when re-recorded
under a new run id with a sealed cross-family review. Those three fields return to main's bytes; the premise
correction and the codex_tool arm are carried forward as an open item for that re-record (decision record).
agent-sdks next_gap now places the OpenAI Agents SDK as an arm of the layer's single acceptance gate (the
preregistered three-arm workers comparison) and states that a promotion path first needs the frozen rows re-recorded
(Codex review P2). catalogs/landscape/foundation.json#/layers/16/candidates is registered as a research supplement
and the decision index regenerated (Codex review P2).

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Re-register the catalog drift files (hot-file protocol, last commit)

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Catalog drift review round: authentication claims, carried #15167 head, stale citations, one codex_tool acceptance path

- Corrects the authentication claims and states the carried OmniRoute #15167 head that 20128's build cherry-picked
  (f5d8e150), apart from the PR's later force-pushed head.
- Refreshes stale citations to pinned URLs, so a reviewer with network can re-read each one.
- agent-sdks next_gap and the decision record name one acceptance path for codex_tool: a newly preregistered comparison
  with the Codex thread API and `codex exec --json` plus resume.
- Minor findings from the Opus and GPT-6.1 Sol reads; no field feeding the frozen wave-20260922 verdict rows changes
  (frozen verdict gate: 0 changed rows, 0 violations).

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Re-register the catalog drift review-round files (hot-file protocol)

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Regenerate the new-WSL handbook for the research-state.json change and refresh its receipt's two output hashes

build_new_wsl_handbook.py --write changes only the research-state.json input-hash line in docs/new-wsl-handbook.md
and .json. No tool writes evidence/artifacts/new-wsl-handbook-20261001/receipt.json, so its two outputs[] hash fields
were edited by hand to the regenerated files' sha256 (nothing else in the receipt changed). Done with the
configuration owner's ACK (option 1).

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Re-register the regenerated handbook and its receipt (hot-file protocol, last commit)

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Set manifests/evidence.json to main 1f5a791's copy (hot-file protocol, step 1 of 2)

The next commit restores this branch's registry, so the registry changes in the branch's last commit.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Re-register the catalog drift files on main 1f5a791 (hot-file protocol, last commit)

Main's registry plus the ten owned bindings; docs/harness-defaults.md carries the resolved file's hash (main's #648 row plus #641's two rows). Foreign rows equal main's, in main's order.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Set manifests/evidence.json to main 1a64e8f's copy (hot-file protocol, step 1 of 2)

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Re-register the catalog drift files on main 1a64e8f (hot-file protocol, last commit)

Main's registry plus the ten owned bindings; docs/harness-defaults.md carries the merged file's hash (main's three 2026-10-03 rows plus #641's two rows at their accepted position), and the two new-WSL handbook outputs and their receipt carry the regenerated and hand-refreshed hashes.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Hot-file protocol: reset manifests/evidence.json to the merge base before the final registry commit

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Register this branch's files on the merge-base registry (hot-file protocol: every registry edit in the last commit)

Co-Authored-By: Claude Opus 5.5 <[email protected]>

---------

Co-authored-by: Scout <scout@local>
Co-authored-by: Claude Opus 5.5 <[email protected]>

…the wave-4 promptfoo owner default on the repository-quality rule

Integrates the eight GPT fix-wave builders (bounded job 050, GPT-6.1 Sol at
max), working from the verified-E2E fix list. Two review rounds followed:
- A five-way Claude read. Its 4 P1 authority wordings were fixed by the
  coordinator, and its 19 P2 recipe defects by GPT repair job 058.
- A Claude delta read. Its P2 (the gateway canary check's missing argument)
  and the leftover wording were fixed by the coordinator.

Result:
- 21 slots are repaired in the install plan, its acceptance and the client
  wiring.
- 4 still need a native sign-in or a destination: alerting, grafana,
  promptfoo and ccusage.
- OpenHands stays at v1.50.1, the manifest baseline.
- Every revised destination command is UNRUN here; the co-op session
  applies them on NativeStack2604.

wave4: an owner default on promptfoo, resting on the owner's 2026-10-04
repository-quality rule. Decide round 2 and its Opus verification named
promptfoo for paired skill verification in both clients
(docs/decisions/2026-10-04-repository-quality-rule.md;
evidence/artifacts/new-wsl-fix-wave-20261004/).

CI repair (the coordinator, after the first push):
- The shared top-rule templates go back to main: codex.AGENTS.template.md
  and examples/claude-native/CLAUDE.md (the 8 KiB budget and the verbatim
  surfaces).
- The host stack manifest and its freshness snapshot go back to main. A
  host pin move needs its saturation-audit row and a qualification receipt
  (tests/test_stack_lifecycle.py:21-37), so srt, Worktrunk, Collector
  Contrib, Grafana and Syft are left to the currency PRs.
- The 2604 plan, profile and architecture keep the plan's pins, and
  build_ecosystem.py --check labels the drift.
- Every architecture pin_source cites its component's version line in the
  current stack, including two that #637 had shifted on main.

Rebuilt on main 5df0e0e (after #699, #700 and #637) as one net content
commit. Generated outputs were regenerated with their own tools, and the
evidence registry follows in the last commit. The earlier commits stay
reachable as 1572f48, 89d7a2c and 8ca0d7a for the review record.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

… citation and receipt fixes (Opus read of merges 2+3)

- docs/decisions/2026-10-03-currency-wave-w1.md has a dated section, "Merge with main e871259
  and 4c89741 (2026-10-05)". It records:
  - the code-index owner rows aligned to jcodemunch-mcp 1.108.327, which
    test_owner_pins_are_the_repository_pins requires;
  - every configuration-owner file this PR now edits;
  - the snapshot counters after main's #637 recount (264 + 3 = 267; 281 + 3 = 284);
  - the two new handbook receipt entries.

  The earlier handoff text is kept as dated history. The current handoff leaves only
  playwright-cli 0.1.21 open (stack 0.1.22); Syft 1.54.0 and mcporter 0.14.2 now match.
- The handbook receipt's new entries carry their own UTC dates and a recorded --check run.
- install-plan SOURCES.md cites recipes/README.md at the merged line numbers. install.sh again
  carries a second source for --python 3.13 and the ecosystem root. The Wave 3 intros name the
  jcodemunch-mcp 1.108.327 move (source 6d5ae86c) and its dated README read.
- docs/foundation-stack.md names the sixth predated component, jCodeMunch 1.108.319.

Built by GPT-6.1 Sol through the packaged SDK worker (round 642-merge-fix); committed by the
coordinator.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

Resolve the browser and gateway conflicts against the fixed round-2 owners.
Wire Chrome DevTools MCP 1.10.1 through the destination client map in both
native clients, retaining the host agent-browser reference and the host stack
bytes. Select published OmniRoute 3.8.51, preserve the tool_root composition
binding for carried prefixes, and require the selected release identity.
Integrate Harbor telemetry, Promptfoo skill verification and research routing.

Regenerate the owner inventory and shared projections with their original
tools. Update strict wave-5 tests from the recorded decision/consensus, retain
prior decisions under overturned, and record every conflict or deviation in
integration-resolutions.json.

SOTA sources:
- this PR:docs/decisions/2026-10-04-final-architecture-round2.md:23-38
- this PR:evidence/artifacts/new-wsl-layer-consensus-20261002/consensus.json:1308
- main PR #637:docs/decisions/2026-10-03-omniroute-3851-pin.md:108
- ChromeDevTools/chrome-devtools-mcp@e52c6b5:docs/client-configurations.md:71,109
- cheaper-inference/OmniRoute@c1e30b7:bin/cli/commands/doctor.mjs:632
- harbor-framework/harbor@1e5c5c6:tests/unit/test_trajectory_validator.py:1

Validation:
- check_plan: 84 rows, 63 selected (60 default, 3 named), 2 measurement-only,
  19 excluded, 186 commands, 111 acceptance entries; zero problems.
- Shell syntax, client-map, handbook and ecosystem checks pass; the ecosystem
  reports 23 preserved historical pin drifts.
- Required seven-module unittest suite: 542 tests, 536 passed, 6 skipped.
- validate.py: exit 1 with registry drift only; registry reconciliation is
  coordinator-owned. git diff --check passes.

No destination install, upstream runtime acceptance or model/GPU run is
claimed. The shared gateway effort comment, host-currency moves, OpenHands
telemetry producer reconciliation and evidence registry remain with their
owners. Working-tree edits only; the coordinator stages and commits.

Integrated by GPT bounded job r2-integrate (GPT-6.1 Sol at max) on the coordinator-applied G1/G3 commit and the G2/G4 patches; the coordinator commits and registers.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

Resolve the browser and gateway conflicts against the fixed round-2 owners.
Wire Chrome DevTools MCP 1.10.1 through the destination client map in both
native clients, retaining the host agent-browser reference and the host stack
bytes. Select published OmniRoute 3.8.51, preserve the tool_root composition
binding for carried prefixes, and require the selected release identity.
Integrate Harbor telemetry, Promptfoo skill verification and research routing.

Regenerate the owner inventory and shared projections with their original
tools. Update strict wave-5 tests from the recorded decision/consensus, retain
prior decisions under overturned, and record every conflict or deviation in
integration-resolutions.json.

SOTA sources:
- this PR:docs/decisions/2026-10-04-final-architecture-round2.md:23-38
- this PR:evidence/artifacts/new-wsl-layer-consensus-20261002/consensus.json:1308
- main PR #637:docs/decisions/2026-10-03-omniroute-3851-pin.md:108
- ChromeDevTools/chrome-devtools-mcp@e52c6b5:docs/client-configurations.md:71,109
- cheaper-inference/OmniRoute@c1e30b7:bin/cli/commands/doctor.mjs:632
- harbor-framework/harbor@1e5c5c6:tests/unit/test_trajectory_validator.py:1

Validation:
- check_plan: 84 rows, 63 selected (60 default, 3 named), 2 measurement-only,
  19 excluded, 186 commands, 111 acceptance entries; zero problems.
- Shell syntax, client-map, handbook and ecosystem checks pass; the ecosystem
  reports 23 preserved historical pin drifts.
- Required seven-module unittest suite: 542 tests, 536 passed, 6 skipped.
- validate.py: exit 1 with registry drift only; registry reconciliation is
  coordinator-owned. git diff --check passes.

No destination install, upstream runtime acceptance or model/GPU run is
claimed. The shared gateway effort comment, host-currency moves, OpenHands
telemetry producer reconciliation and evidence registry remain with their
owners. Working-tree edits only; the coordinator stages and commits.

Integrated by GPT bounded job r2-integrate (GPT-6.1 Sol at max) on the coordinator-applied G1/G3 commit and the G2/G4 patches; the coordinator commits and registers.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

Resolve the browser and gateway conflicts against the fixed round-2 owners.
Wire Chrome DevTools MCP 1.10.1 through the destination client map in both
native clients, retaining the host agent-browser reference and the host stack
bytes. Select published OmniRoute 3.8.51, preserve the tool_root composition
binding for carried prefixes, and require the selected release identity.
Integrate Harbor telemetry, Promptfoo skill verification and research routing.

Regenerate the owner inventory and shared projections with their original
tools. Update strict wave-5 tests from the recorded decision/consensus, retain
prior decisions under overturned, and record every conflict or deviation in
integration-resolutions.json.

SOTA sources:
- this PR:docs/decisions/2026-10-04-final-architecture-round2.md:23-38
- this PR:evidence/artifacts/new-wsl-layer-consensus-20261002/consensus.json:1308
- main PR #637:docs/decisions/2026-10-03-omniroute-3851-pin.md:108
- ChromeDevTools/chrome-devtools-mcp@e52c6b5:docs/client-configurations.md:71,109
- cheaper-inference/OmniRoute@c1e30b7:bin/cli/commands/doctor.mjs:632
- harbor-framework/harbor@1e5c5c6:tests/unit/test_trajectory_validator.py:1

Validation:
- check_plan: 84 rows, 63 selected (60 default, 3 named), 2 measurement-only,
  19 excluded, 186 commands, 111 acceptance entries; zero problems.
- Shell syntax, client-map, handbook and ecosystem checks pass; the ecosystem
  reports 23 preserved historical pin drifts.
- Required seven-module unittest suite: 542 tests, 536 passed, 6 skipped.
- validate.py: exit 1 with registry drift only; registry reconciliation is
  coordinator-owned. git diff --check passes.

No destination install, upstream runtime acceptance or model/GPU run is
claimed. The shared gateway effort comment, host-currency moves, OpenHands
telemetry producer reconciliation and evidence registry remain with their
owners. Working-tree edits only; the coordinator stages and commits.

Integrated by GPT bounded job r2-integrate (GPT-6.1 Sol at max) on the coordinator-applied G1/G3 commit and the G2/G4 patches; the coordinator commits and registers.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

Resolve the browser and gateway conflicts against the fixed round-2 owners.
Wire Chrome DevTools MCP 1.10.1 through the destination client map in both
native clients, retaining the host agent-browser reference and the host stack
bytes. Select published OmniRoute 3.8.51, preserve the tool_root composition
binding for carried prefixes, and require the selected release identity.
Integrate Harbor telemetry, Promptfoo skill verification and research routing.

Regenerate the owner inventory and shared projections with their original
tools. Update strict wave-5 tests from the recorded decision/consensus, retain
prior decisions under overturned, and record every conflict or deviation in
integration-resolutions.json.

SOTA sources:
- this PR:docs/decisions/2026-10-04-final-architecture-round2.md:23-38
- this PR:evidence/artifacts/new-wsl-layer-consensus-20261002/consensus.json:1308
- main PR #637:docs/decisions/2026-10-03-omniroute-3851-pin.md:108
- ChromeDevTools/chrome-devtools-mcp@e52c6b5:docs/client-configurations.md:71,109
- cheaper-inference/OmniRoute@c1e30b7:bin/cli/commands/doctor.mjs:632
- harbor-framework/harbor@1e5c5c6:tests/unit/test_trajectory_validator.py:1

Validation:
- check_plan: 84 rows, 63 selected (60 default, 3 named), 2 measurement-only,
  19 excluded, 186 commands, 111 acceptance entries; zero problems.
- Shell syntax, client-map, handbook and ecosystem checks pass; the ecosystem
  reports 23 preserved historical pin drifts.
- Required seven-module unittest suite: 542 tests, 536 passed, 6 skipped.
- validate.py: exit 1 with registry drift only; registry reconciliation is
  coordinator-owned. git diff --check passes.

No destination install, upstream runtime acceptance or model/GPU run is
claimed. The shared gateway effort comment, host-currency moves, OpenHands
telemetry producer reconciliation and evidence registry remain with their
owners. Working-tree edits only; the coordinator stages and commits.

Integrated by GPT bounded job r2-integrate (GPT-6.1 Sol at max) on the coordinator-applied G1/G3 commit and the G2/G4 patches; the coordinator commits and registers.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

Record a dated compatibility amendment for the retained session-80 profile
and #637's source-backed OmniRoute 3.8.51 route. Bind tests to the baseline
and exact configuration hashes without claiming a new measurement. Update
the static header and preflight assertion to the already-correct
per-run x-omniroute-session-id header.

Use grep -E for only the Chrome Signed-By and Claude MCP registration
checks, preserving their patterns and existing negative controls; synchronize
install-plan.json. Retain all other rg uses and previous repair behavior.

Reproduce and verify the four CI failures with rg absent from PATH.

Validation with rg absent from PATH: all six requested test modules passed
(469 tests), check_plan reported 0 problems, and handbook and accept.sh syntax
checks passed. validate reports registry drift only; evidence.json and the
other protected files and git index remain unchanged.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

Resolve the browser and gateway conflicts against the fixed round-2 owners.
Wire Chrome DevTools MCP 1.10.1 through the destination client map in both
native clients, retaining the host agent-browser reference and the host stack
bytes. Select published OmniRoute 3.8.51, preserve the tool_root composition
binding for carried prefixes, and require the selected release identity.
Integrate Harbor telemetry, Promptfoo skill verification and research routing.

Regenerate the owner inventory and shared projections with their original
tools. Update strict wave-5 tests from the recorded decision/consensus, retain
prior decisions under overturned, and record every conflict or deviation in
integration-resolutions.json.

SOTA sources:
- this PR:docs/decisions/2026-10-04-final-architecture-round2.md:23-38
- this PR:evidence/artifacts/new-wsl-layer-consensus-20261002/consensus.json:1308
- main PR #637:docs/decisions/2026-10-03-omniroute-3851-pin.md:108
- ChromeDevTools/chrome-devtools-mcp@e52c6b5:docs/client-configurations.md:71,109
- cheaper-inference/OmniRoute@c1e30b7:bin/cli/commands/doctor.mjs:632
- harbor-framework/harbor@1e5c5c6:tests/unit/test_trajectory_validator.py:1

Validation:
- check_plan: 84 rows, 63 selected (60 default, 3 named), 2 measurement-only,
  19 excluded, 186 commands, 111 acceptance entries; zero problems.
- Shell syntax, client-map, handbook and ecosystem checks pass; the ecosystem
  reports 23 preserved historical pin drifts.
- Required seven-module unittest suite: 542 tests, 536 passed, 6 skipped.
- validate.py: exit 1 with registry drift only; registry reconciliation is
  coordinator-owned. git diff --check passes.

No destination install, upstream runtime acceptance or model/GPU run is
claimed. The shared gateway effort comment, host-currency moves, OpenHands
telemetry producer reconciliation and evidence registry remain with their
owners. Working-tree edits only; the coordinator stages and commits.

Integrated by GPT bounded job r2-integrate (GPT-6.1 Sol at max) on the coordinator-applied G1/G3 commit and the G2/G4 patches; the coordinator commits and registers.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

Record a dated compatibility amendment for the retained session-80 profile
and #637's source-backed OmniRoute 3.8.51 route. Bind tests to the baseline
and exact configuration hashes without claiming a new measurement. Update
the static header and preflight assertion to the already-correct
per-run x-omniroute-session-id header.

Use grep -E for only the Chrome Signed-By and Claude MCP registration
checks, preserving their patterns and existing negative controls; synchronize
install-plan.json. Retain all other rg uses and previous repair behavior.

Reproduce and verify the four CI failures with rg absent from PATH.

Validation with rg absent from PATH: all six requested test modules passed
(469 tests), check_plan reported 0 problems, and handbook and accept.sh syntax
checks passed. validate reports registry drift only; evidence.json and the
other protected files and git index remain unchanged.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

… rule, and NativeStack2604 plan rows for hcom, Chrome DevTools MCP, DSPy, SkillSpector, Inspect Scout and MCP conformance (#713)

### Scope

- What this PR changes: the final-architecture round 2 (wave 5). It records decide round 2's owners as consensus batch `wave5`, on the owner's 2026-10-04 repository-quality rule, and adds NativeStack2604 install-plan rows for them:
  - hcom 0.7.27 for Claude Code ↔ Codex messaging;
  - Chrome DevTools MCP 1.10.1 for browser automation and diagnostics (`web-research/playwright-cli`);
  - DSPy 3.4.0 with GEPA 0.1.4 for LM program optimization;
  - SkillSpector 2.12.0 for skill vetting;
  - Inspect Scout 0.5.3 for trajectory analysis;
  - the MCP conformance suite;
  - configure-only rows for promptfoo skill-lifecycle verification, the research-service LLM route, the gpt-gateway topology (OmniRoute 3.8.51, following #637) and Harbor worker telemetry.
- Base commit: `4c897418f` (#704 landed)
- Lane: `lane:foundation`
- Owned paths touched:
  - `evidence/artifacts/new-wsl-install-plan-20261002/**`;
  - `adoption/new-wsl/client-config-map.json` and its templates;
  - `adoption/new-wsl-profile.{json,md}`;
  - `catalogs/foundation/new-wsl-architecture-20261001.json`;
  - the consensus and definitive-manifest records;
  - `docs/decisions/2026-10-04-final-architecture-round2.md` and the four `2026-10-04-round2-plan-g*.md` records;
  - the new-WSL tests;
  - the generated handbook.

`manifests/stack.json` and `catalogs/landscape/upstream-snapshot.json` are unchanged. Host-stack moves need qualification receipts and belong to currency PRs.

### SOTA sources

- aannoo/hcom tag v0.7.27 = 2c5f343b2f9ec4bf2acf49c0431860e7c2ae578b (installer checksum in the G1 record).
- ChromeDevTools/chrome-devtools-mcp tag chrome-devtools-mcp-v1.10.1 = e52c6b59b476c5e04d8dd9fd4bd017ba3b3d65df (npm `[email protected]`).
- stanfordnlp/dspy tag 3.4.0 = 2413b67a4d08a476e4bc6f40b9f8f42f87711ee7 (`dspy==3.4.0`; GEPA 0.1.4 as its declared dependency).
- NVIDIA/SkillSpector v2.12.0 = c7958a3268d9498644b22edb75d0f051bbc8cbfc (release-asset sha256 in the G3 record).
- meridianlabs-ai/inspect_scout tag 0.5.3 = 0e8fc055a3cebba1a14c11bc35856767b6405173.
- modelcontextprotocol/conformance: `@modelcontextprotocol/conformance` 0.2.0-alpha line (the G3 record states the pinned alpha and its cooldown).
- The repository-quality rule and each verdict's install and acceptance notes: `docs/decisions/2026-10-04-repository-quality-rule.md` and `evidence/artifacts/final-architecture-round2-20261004/verdicts.json`. Every row cites its upstream file:line in `SOURCES.md` and in the group records.

### Evidence-class table

| Claim | Evidence class | Command / receipt |
| --- | --- | --- |
| The plan's scripts and rows agree, and every owner row installs its default | local_integration | `check_plan.py`: 0 problems [integrator] |
| The client map renders, and the handbook and receipts are regenerated | local_integration | `new_wsl_client_config.py --check` 0, `build_new_wsl_handbook.py --check` 0 [integrator, coordinator] |
| The architecture builds; the stack drift is labelled, not failing | local_integration | `build_ecosystem.py --check` 0, with 23 drift cells listed [integrator] |
| The destination installs and their upstream acceptance | none yet (UNRUN) | applied by the NativeStack2604 co-op session after landing |

### Local commands run

```
$ python3 -m unittest tests.test_new_wsl_definitive_defaults tests.test_new_wsl_client_config tests.test_new_wsl_handbook tests.test_new_wsl_profile tests.test_stack_lifecycle tests.test_ecosystem_manifest tests.test_upstream_surface_watch
542 tests, 536 passed, 6 skipped   [integrator]
$ python3 scripts/validate.py
status passed (10,103 hashed files, 202 receipts)   [coordinator, after the registry commit]
$ merge_tree_landing_check.py origin/main HEAD
LANDABLE
```

### Decision record

- `docs/decisions/2026-10-04-final-architecture-round2.md` (the owners, the rule, and the overturn conditions);
- the group records: `docs/decisions/2026-10-04-round2-plan-g1-messaging.md`, `-g2-browser.md`, `-g3-tools.md` and `-g4-config.md`;
- the integration ledger `evidence/artifacts/final-architecture-round2-20261004/integration-resolutions.json`.

Also included: the qualifying note on the fix-wave citation ledger's superseded stack rows, which was the residual P3 of #704's GPT delta read (job-065).

Built by four GPT builder jobs and a GPT integrator (GPT-6.1 Sol at max). The coordinator applied the patches, committed, and registered. The cross-family Claude read is pending, so this PR stays a draft until it returns.




### Review chain and residuals (bounded loop)

- **Claude four-lens read:** 3 P1, 12 P2, 24 P3.
- **GPT repair round 1:** fixed all P1 and P2.
- **Claude delta read:** 0 P1, 3 P2, 5 P3.
- **GPT repair round 2:** fixed N1-N8, R2-19, R2-22 and R2-36.
- **Claude re-check at 7fbe586:** 0 P1, 0 P2, 4 P3.

**Residuals, routed to the post-landing drift-repair PR:**
1. The gateway-effort check sends its stdout to /dev/null at `P/accept.sh:61`, so a pass via the time-window fallback looks the same as a pass on the session match. Its embeddings rejection can never fire, because embeddings rows carry no session tag.
   - Fix: write the join mode and caveats to stderr; return 78 on the fallback; re-scope the embeddings claim.
2. The Chrome origin check relies on `apt-cache madison`. After the next Chrome release it fails until Chrome is upgraded.
   - Fix: take the origin from the native source file plus the `apt-cache policy` candidate, and report an older-than-candidate install as 78; add a stale-index test.
3. `adoption/new-wsl-profile.json:1026` still names the exact Chrome build.
   - Fix: "current stable, minimum 154.0.8037.97-1", then regenerate the handbook.
4. One stale effort claim remains, in `integration-resolutions.json` `repair_round.completeness_critic`: it needs an `amended_by` note.

**Carried forward** (owners named in the ledger): R2-20 and R2-27.

**Verdict files**, in the coordinator's bounded-job folders:
- verdict-claude-713-delta.md
- verdict-claude-713-recheck.md

### Recorded residuals (bounded review, 2026-10-05)

Claude Opus gate read of round 3 at `cde8a0de060c` returned ACCEPT with no p1. Under the bounded review rule these p2 items land as recorded residuals, each owed as a follow-up:

- `evidence/artifacts/new-wsl-install-plan-20261002/README.md:279-283 and install-plan.json:2658 (alerting row needs_user)`: Round 3 dropped main's and the accepted head's note that ntfy.sh alerts leave the host ('recommended;

### Recorded residuals (bounded review, 2026-10-05; Claude Opus 5.5 gate read (session 5f) at 6eb8b94 at 6eb8b94)

Claude Opus 5.5 gate read (session 5f) at 6eb8b94 at `6eb8b9407092` returned ACCEPT with no p1. Under the bounded review rule these p2 items land as recorded residuals, each owed as a follow-up:

- `evidence/artifacts/new-wsl-install-plan-20261002/install-plan.json:4215 (owners[77] 'GPT Researcher and DeerFlow', slot research-harnesses, `: The note still calls the committed config 'the measured session-80 values without the key: ...
- `evidence/artifacts/new-wsl-install-plan-20261002/README.md:279-283 and install-plan.json:2619 (owners[45] Alertmanager, slot alerting, field`: Carried from r3, still not fixed.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

…Sol max (published v3.8.51 + #15167 + the cited affinity patch) (#744)

### Scope

- What this PR changes, in one or two sentences: records the decision that the NativeStack OmniRoute gateways keep true GPT-6.1 Sol max (published `v3.8.51` content plus upstream PR 15167 plus the cited local session-affinity patch) instead of switching to the published-only package, with the wire evidence, the affinity commit as a public patch file (sha256) and the build facts of the composition. Nothing running changes: the planned 15:00Z swap was cancelled before any unit changed. The second host's claim is scoped to what is observed: NativeStack 20128 delivers max; on 2604 the co-op's read-back of 16:50Z (relayed, private record) gave HTTP 200 on the `-max`, `-xhigh`, bare and `astra-max` routes and upstream `max` on `sol-max` for 21128; the 2604 plan rows, which since #713 select the published 3.8.51 with Sol at xhigh, are handed to the plan owner (the fixwave-defects lane holds the plan files through #723), not edited here. Standing constraint F1 is extended to 21128/21129.
- Base commit: `cb339488e`, merged with main `2d849ba1f` (head `08e749ec4`, after the repair rounds for 5f's read of `b5bf717d9`, its delta reads of `fb3d1ee40` and `3d65b80fa` and c5's read of `7200ee424`; main moved with #713, so the plan rows the record cites are the post-#713 ones)
- Lane: `lane:shared` (the gateway serves the foundation and trading lanes, as in #637); the trading lane's acknowledgement is requested before merge ([docs/lanes.md](../docs/lanes.md)).
- Owned paths touched: `docs/decisions/2026-10-05-omniroute-gateway-composition.md`, `evidence/artifacts/omniroute-wire-effort-20261005/` (new), one-clause F1 extensions in `docs/decisions/2026-09-30-omniroute-rebuild.md` and `docs/foundation-stack.md`, and the hot-file registry rows in `manifests/evidence.json` (last commit, hot-file protocol).

### SOTA sources

- OmniRoute upstream, https://github.com/diegosouzapw/OmniRoute: tag `v3.8.51` (commit `c1e30b7676975feb298b49eff6ff58923c04b89e`, tree `0f58d8df20c0c2ae4336b432b3f39837119b6eed`, equal to the tree of `release/v3.8.51` at `2f42a9ac19d1a247ec9ce5473b790843724b3061`); npm `[email protected]` (integrity `sha512-VwwSt+bP9lJiPJXFJMz0nNGGuoewPZU3nFe1SLuO11ADgdSwTegGCxhg8Ov75+31m/cocPxHiO63zygn1XQ0MQ==`).
- Upstream PR 15167 (head `0585aba5589d5a1f49243a13a8db249558e7c9e3`): `gpt-6.1-sol` in the Codex registry and alias sets; its re-pin trigger is the first official release that carries it. Upstream issue 8939 and PR 8940 (the affinity premise), open PR 13102, PR 13788 (commits `24bbadbad15eec0d598c908416e4b01be4e37671` and `6c7990058c4ce9677de79452c8cefb10b4bf1b3d`; optional, dropped at the next rebuild). The running builds carry PR 15167's earlier head `f5d8e150b` and local cherry-picks of PR 13788's two commits with the same patch-ids.
- Upstream code read for the wire behavior: `open-sse/executors/codex.ts` (`clampEffort`, `transformRequest`) and `open-sse/executors/codex/reasoningSuffix.ts` at the tag and with the PR.
- Upstream's own build and checks, unchanged: `npm ci`, `npm run build:release`, `OMNIROUTE_ALLOW_CANARY_BUILD=1 npm run check:pack-artifact`, `npm pack`, `npm run check:pack-boot`.
- The earlier records this one extends: `docs/decisions/2026-09-30-omniroute-rebuild.md`, `docs/decisions/2026-10-03-omniroute-3851-pin.md`, `evidence/artifacts/omniroute-sol-max-20260930/`.

### Evidence-class table

| Claim | Evidence class | Command / receipt |
| --- | --- | --- |
| With PR 15167 the executor sends model `gpt-6.1-sol` and effort `max` for `gpt-6.1-sol-max`, `-ultra` and `gpt-6.1-sol` with client `max`; without it `xhigh`, `medium` and an unsplit model name | `local_integration` (upstream code, our probe; nothing was sent) | `checks/wire-probe-with-15167.txt`, `checks/wire-probe-without-15167.txt` |
| The live call log shows requested `max` and upstream `max` for Sol after the 2026-09-30 restart (7,550 and 2,728 rows) and `xhigh` before it (1,165 rows) | `local_integration` (read-only read of metadata columns; the upstream effort is the one in the provider request the attempt captured, only for rows with encrypted reasoning) | `checks/live-effort.json`, `checks/live-requests.json` |
| The deployed prefix carries the probed code (four files byte-identical, bundle literal counts 40/8 against 0/0 for the clean package) | `local_integration` | `checks/source-identity.json` |
| The base tree equals the tag commit's tree and so the published package content | `source_review` (git facts, npm metadata) | `checks/source-identity.json` |
| The patch file applies to the tag and, with PR 15167's diff, reproduces the trees of the two built commits | `local_integration` (git, temporary index) | `checks/composition-trees.txt`, `patches/045aa81f3.patch` |
| The affinity patch's unit test (7 tests): 6 fail on the tag with only the test file (negative control), 7 of 7 pass on the tag plus the patch and on the composition; the wire probe re-run on both reproduces the committed probe files byte for byte | `local_integration` (upstream runner flags, trees built from public refs, our test and probe) | `checks/affinity-test-results.txt`, `scripts/affinity-test.sh.txt` |
| PR 13788's two public commits have the same patch-ids as the four local cherry-picks; the carry bundle restores into a fresh repo | `local_integration` (git) | `checks/pr13788-carried-patch-ids.txt` |
| The composition builds with upstream's release scripts and its tarball boots and persists in a clean prefix | `local_integration` (upstream checks on our build; the first `check:pack-boot` run failed on a port collision and passed on the rerun, both recorded; the kept tarball equals a fresh `npm pack --dry-run` of the booted tree by sha1 and sha512) | `checks/build-notes.txt` |
| 7-day use of `/v1/responses`: 20,371 calls, 93.2% cache reads; 21.7% of calls use `-max`/`-ultra` names | `local_integration` (read-only counts) | `checks/usage-7d.json` |
| Without #15167 the live gateway rejects `cx/gpt-6.1-sol-max` and `-xhigh` at the catalog gate (HTTP 400) and clamps `cx/gpt-6.1-sol` with `max` to `xhigh`; with it all answer 200 with upstream `max`, `max`, `xhigh`, `max` | `local_integration` (main's recorded live probes of 2026-09-30, cited, not re-run) | `evidence/artifacts/omniroute-sol-max-20260930/checks/probe-gate-before-switch.json` (C7), `probe-gate-after-switch.json` |
| The kept tarball carries PR 15167's head of 2026-10-05 (17 files), not the running build's head of 2026-09-30 (5 files); the 13 test files covering the extra pieces pass on the composition, 145 of 145 | `local_integration` (upstream test runner on a tree built from public refs); the version-header change is not observed live | `checks/affinity-test-results.txt` (last block), `evidence/artifacts/omniroute-sol-max-20260930/checks/upstream-pr-15167-identity.txt` |
| NativeStack2604's 21128 delivers max on `sol-max`; `-xhigh`, bare and `astra-max` answer 200 | relayed from the co-op's read-back (an exact `X-Correlation-Id` join; request headers unobservable); the files are in its private coordination record, not in the repository | the co-op's ledger row 165024Z, cited as relayed by the command center |
| The backend applies `max` | not observable from our side; not claimed | |
| The cache cost of dropping the affinity patch | not measured | |

### Local commands run

```
$ python3 scripts/validate.py
{"components": 69, "hashed_files": 10232, "profiles": 4, "receipts": 204, "status": "passed"}
$ python3 scripts/validate_convergence.py --all-recorded
record[30]: valid
$ python3 -m unittest tests.test_component_matrix tests.test_evidence_manifest tests.test_new_host_grand_list tests.test_validate tests.test_frozen_macos_variant_no_use
Ran 172 tests ... OK; Ran 10 tests ... OK
(gitleaks, run by the commit hook on each commit: no leaks found)
```

### Decision record

`docs/decisions/2026-10-05-omniroute-gateway-composition.md`: the composition table with each carried change at its commit and its removal or re-pin condition, the user's choices as the command center relayed them, the swap-impact measurements, the wire evidence, alternatives A/B/C, limits and the overturn conditions.

### Host evidence

Not applicable (no files under `evidence/hosts/`).

### Checklist

- [x] New/changed GitHub Actions are pinned to a full commit SHA with a version comment (no workflow changed).
- [x] New/changed workflows declare top-level `permissions: {}` (no workflow changed).
- [x] No secrets are printed, logged or committed; no new required secret was added (callers use the documented `local-loopback` placeholder; no connection id, account or request body is recorded).
- [x] No new paid hosting, subscription or billing surface was introduced.
- [x] Peer-owned untracked files and worktrees were preserved.

### Recorded residuals (bounded review, 2026-10-05)

GPT-6.1 Sol micro delta read (session 5f) at 08e749e at `08e749ec4b8c` returned FINDINGS with no p1. Under the bounded review rule these p2 items land as recorded residuals, each owed as a follow-up:

- ``: 
- ``: 

🤖 Generated with [Claude Code](https://claude.com/claude-code)