Skip to content

v0.0.124

# NemoClaw v0.0.124 release brief
@jyaunches jyaunches tagged this 14 Sep 14:18 
## Release range

- Previous release: `v0.0.123` at `f75f722bb4a1ec9642c8df36c8924e24500d78f0`
- Candidate: `6f3cced4230ae9660c049cc11804daf37797c595`
- Candidate selection: current-main
- Commits: 116
- Risky files detected: 388

## QA context

### Risky areas

- Workflow / enforcement
- Sandbox / policy / SSRF
- Credentials / inference
- Installer / bootstrap
- Onboarding / host glue

### Suggested test focus

- Fresh install and upgrade paths
- Onboarding wizard and sandbox creation
- Policy enforcement, network egress, and SSRF protections
- CI checks, pre-commit hooks, and DCO declarations
- Credential storage and inference provider routing

## Release decision status

Prepared for final tag confirmation. No tag has been created. The maintainer requested tagging with the selected E2E evidence and approved the exception rationale below. The full E2E run is still active; its outstanding result is explicitly accepted as unresolved.

## Canonical release entry

- Path: `docs/changelog/2026-09-14.mdx`
- Exactly one matching version heading; detailed bullets verified at the candidate.

## v0.0.124

NemoClaw v0.0.124 updates the managed OpenShell runtime to 0.0.116 and improves upgrades, local inference, and sandbox recovery.
It also restores interactive Deep Agents Code execution features and improves agent message handling.

- Upgrade recovery now waits for the replaced packaged gateway to bind before recovering sandboxes.
  Failed required pre-upgrade backups no longer leave incomplete directories that appear to be recovery points; failed cleanup reports the retained path.
  Rebuild also preserves the recorded serving-profile selection instead of losing it during recreation.
  Related changes: [PR #11251](https://github.com/NVIDIA/NemoClaw/pull/11251), [PR #11710](https://github.com/NVIDIA/NemoClaw/pull/11710), [PR #11589](https://github.com/NVIDIA/NemoClaw/pull/11589), and [PR #11686](https://github.com/NVIDIA/NemoClaw/pull/11686).
- Named sandbox commands now use the sandbox's recorded gateway across gateway-port registry roots.
  Portable Hermes recovery reconciles a saved stopped OpenShell state and repairs required forwards.
  If forward startup fails, recovery removes only forwards it started and reports cleanup it cannot verify.
  Interactive `launch` releases lifecycle locks after dispatch so other lifecycle commands can proceed during the session.
  Related changes: [PR #11445](https://github.com/NVIDIA/NemoClaw/pull/11445), [PR #11668](https://github.com/NVIDIA/NemoClaw/pull/11668), [PR #11670](https://github.com/NVIDIA/NemoClaw/pull/11670), and [PR #11657](https://github.com/NVIDIA/NemoClaw/pull/11657).
- llama.cpp selection uses the authenticated server's served context window when no valid explicit override is set.
  Docker Desktop on WSL checks the host-loopback listener from the WSL CLI process before the sandbox bridge check.
  Unsupported inference-provider errors list supported providers on the selected gateway, with a supported-provider fallback when discovery fails.
  Related changes: [PR #11544](https://github.com/NVIDIA/NemoClaw/pull/11544), [PR #11444](https://github.com/NVIDIA/NemoClaw/pull/11444), and [PR #11464](https://github.com/NVIDIA/NemoClaw/pull/11464).
- Interrupted managed vLLM installation can replace a running container only when its ownership, local Docker connection, and port mapping match the requested installation.
  Other port holders remain protected.
  N1x preview rebuild accepts a retained canonical managed endpoint with the recorded preview decision.
  Related changes: [PR #11587](https://github.com/NVIDIA/NemoClaw/pull/11587) and [PR #11580](https://github.com/NVIDIA/NemoClaw/pull/11580).
- Native Linux Docker OpenClaw configuration export accepts the verified Ollama route created during ordinary onboarding.
  It requires the managed proxy credential and omits its value from the export.
  Stopping an Ollama sandbox no longer retains its model solely because a matching sibling is stopped or incomplete.
  An unverified sibling state prevents model unload and makes `stop` exit nonzero.
  Related changes: [PR #11550](https://github.com/NVIDIA/NemoClaw/pull/11550) and [PR #11656](https://github.com/NVIDIA/NemoClaw/pull/11656).
- Managed Deep Agents Code interactive sessions restore native local shell, startup-command, and interpreter behavior inside the OpenShell sandbox.
  Headless shell execution remains disabled, and headless non-shell tools retain automatic approval.
  Non-interactive reply handling and pre-session inference verification also improve.
  Related changes: [PR #11683](https://github.com/NVIDIA/NemoClaw/pull/11683), [PR #11276](https://github.com/NVIDIA/NemoClaw/pull/11276), and [PR #11546](https://github.com/NVIDIA/NemoClaw/pull/11546).
- Agent message dispatch preserves piped input and avoids waiting on idle standard input when a message is already supplied.
  Managed OpenClaw heartbeats use isolated sessions to avoid changing the main conversation.
  Related changes: [PR #11382](https://github.com/NVIDIA/NemoClaw/pull/11382) and [PR #11487](https://github.com/NVIDIA/NemoClaw/pull/11487).
- Fresh interceptor-based providerless onboarding now leaves OpenClaw and Hermes without an inference route until a managed provider is configured.
  Version 2 external-component declarations remain unqualified against a released OpenShell runtime and must not be relied on in an installed release.
  Related changes: [PR #11584](https://github.com/NVIDIA/NemoClaw/pull/11584), [PR #11633](https://github.com/NVIDIA/NemoClaw/pull/11633), and [PR #11525](https://github.com/NVIDIA/NemoClaw/pull/11525).
- Discord setup rejects the literal bot-token placeholder before configuring the channel.
  Uninstall failures caused by unavailable Docker now explain how to restore access and retry while preserving gateway state.
  Fatal messaging-build and npm remediation diagnostics redact sensitive process details.
  Related changes: [PR #11404](https://github.com/NVIDIA/NemoClaw/pull/11404), [PR #11474](https://github.com/NVIDIA/NemoClaw/pull/11474), and [PR #11674](https://github.com/NVIDIA/NemoClaw/pull/11674).

## Documentation coverage

- Latest included cumulative docs PR: [#11538](https://github.com/NVIDIA/NemoClaw/pull/11538), merged September 14, 2026 at 13:38:26 UTC.
- Final PR commit: `9876e0bf1c9566786fa6bc2654f326a34850f89e`.
- Merge commit: `6f3cced4230ae9660c049cc11804daf37797c595`, equal to the planned candidate.
- Final automated refresh: `c630d2ce5c54f6d5dcc868a23b99bd0cde1a0519`.
- Automated coverage commit: `04966534817cad2c1f89e9db8caefa0a1d6a2a0b`, verified from the workflow-created commit parents.
- Later first-parent commits: only the documentation merge #11538; no uncovered product merge.
- Changed paths: all 16 files are under `docs/`; the full path list appears below.
- Open managed docs PRs: none at inspection.
- Review: GitHub reports `REVIEW_REQUIRED` after merge; do not infer an approval. Independent writer review of the final PR commit found no actionable findings. All 29 PR commits were GitHub Verified before merge.
- Local validation: `npm run docs` passed with zero errors and two warnings (unauthenticated redirect check and existing theme contrast). Normal commit and publication hooks passed.
- Maintainer decision: Proceed with the candidate as shown.
- Decision context: the maintainer merged the reviewed documentation PR, requested preparation to tag, and reiterated that preparation should continue. The only post-coverage commit is that documentation merge.

### Changed documentation paths

- `docs/AUTOMATION.md`
- `docs/changelog/2026-09-14.mdx`
- `docs/deployment/register-external-component.mdx`
- `docs/get-started/prerequisites.mdx`
- `docs/inference/set-up-llama-cpp.mdx`
- `docs/inference/set-up-ollama.mdx`
- `docs/inference/set-up-vllm.mdx`
- `docs/inference/switch-providers.mdx`
- `docs/manage-sandboxes/add-channels-after-onboarding.mdx`
- `docs/manage-sandboxes/enable-channels-during-onboarding.mdx`
- `docs/manage-sandboxes/recover-rebuild-sandboxes.mdx`
- `docs/manage-sandboxes/set-up-discord.mdx`
- `docs/manage-sandboxes/uninstall-nemoclaw.mdx`
- `docs/manage-sandboxes/update-sandboxes.mdx`
- `docs/network-policy/integration-policy-examples.mdx`
- `docs/reference/commands.mdx`

### Documentation PR checks at inspection

| Check | State | Evidence |
|---|---|---|
| commit-lint | CANCELLED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962192/job/103995109885) |
| request | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849956096/job/103995151912) |
| changes | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850013972/job/103995346524) |
| changes | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850615177/job/103997329353) |
| package-openshell-sdk | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850014025/job/103995322914) |
| package-openshell-sdk | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850615046/job/103997329066) |
| assign-linked-issue-author | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850013925/job/103995309302) |
| assign-linked-issue-author | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850614956/job/103997326160) |
| label-release-target | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850493886/job/103996912279) |
| changes | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962139/job/103995110559) |
| Analyze (go) | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849955730/job/103995089753) |
| Docs / Check changed Markdown links | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962233/job/103995109385) |
| Docs / Validate and publish PR preview | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962255/job/103995109864) |
| Docs / Verify CLI and installer reference parity | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962200/job/103995109544) |
| codebase-growth-guardrails | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849958200/job/103995094794) |
| commit-lint | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850014057/job/103995310546) |
| commit-lint | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850615084/job/103997327232) |
| require-maintainer-edits | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849958141/job/103995094900) |
| require-maintainer-edits | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850013884/job/103995285292) |
| require-maintainer-edits | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850614944/job/103997326080) |
| dco-check | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962207/job/103995109856) |
| dco-check | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850013902/job/103995285586) |
| dco-check | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850615153/job/103997326825) |
| CodeQL (javascript-typescript) | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962111/job/103995109670) |
| package-openshell-sdk | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849958146/job/103995096060) |
| check-hash | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962169/job/103995109578) |
| check-hash | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850013949/job/103995285551) |
| check-hash | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850615111/job/103997327293) |
| Analyze (javascript-typescript) | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849955730/job/103995090139) |
| CodeQL (python) | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962111/job/103995109472) |
| Analyze (python) | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849955730/job/103995090099) |
| docs-only-checks | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850013972/job/103995287922) |
| docs-only-checks | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850615177/job/103997330528) |
| docs-only-checks | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962139/job/103995193738) |
| ShellCheck SARIF | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962111/job/103995109375) |
| hugging-face-models | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962139/job/103995196363) |
| hugging-face-models | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850013972/job/103995287310) |
| hugging-face-models | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850615177/job/103997330651) |
| openshell-sdk-package | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962139/job/103995195480) |
| openshell-sdk-package | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850013972/job/103995335257) |
| openshell-sdk-package | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850615177/job/103997330324) |
| reviewed-npm-audit | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962139/job/103995195367) |
| reviewed-npm-audit | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850013972/job/103995287586) |
| reviewed-npm-audit | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850615177/job/103997330393) |
| static-checks | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962139/job/103995196702) |
| static-checks | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850013972/job/103995288762) |
| static-checks | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850615177/job/103997331544) |
| compile-artifacts | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962139/job/103995195750) |
| compile-artifacts | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850013972/job/103995288635) |
| compile-artifacts | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850615177/job/103997334627) |
| installer-integration | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962139/job/103995195821) |
| installer-integration | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850013972/job/103995329679) |
| installer-integration | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850615177/job/103997332566) |
| plugin-tests | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962139/job/103995197052) |
| plugin-tests | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850013972/job/103995288111) |
| plugin-tests | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850615177/job/103997331826) |
| build-typecheck | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962139/job/103995196645) |
| build-typecheck | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850013972/job/103995289290) |
| build-typecheck | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850615177/job/103997333103) |
| cli-test-shards | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962139/job/103995197525) |
| cli-test-shards | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850013972/job/103995288612) |
| cli-test-shards | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850615177/job/103997333358) |
| cli-tests | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962139/job/103995197811) |
| cli-tests | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850013972/job/103995289617) |
| cli-tests | SKIPPED | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850615177/job/103997333871) |
| checks | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34849962139/job/103995715319) |
| checks | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850013972/job/103995287602) |
| checks | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/actions/runs/34850615177/job/103997332749) |
| CodeQL | NEUTRAL | [check](https://github.com/NVIDIA/NemoClaw/runs/103995580611) |
| ShellCheck | SUCCESS | [check](https://github.com/NVIDIA/NemoClaw/runs/103995421892) |
| CodeRabbit | SUCCESS | CodeRabbit skipped bot user |

## Base and managed image evidence

- Base-image candidate: `6f3cced4230ae9660c049cc11804daf37797c595`
- Selected image evidence: [base-image-publication job](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103987429537) succeeded in [full E2E run 34842707308](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308), attempt 2, at `49765a4305ef1f61575f73cf7cb504dbc48d45db`.
- Verified the job name, status, conclusion, SHA, owning run, attempt, workflow path, main branch, and manual event from GitHub API records.
- The repository's `baseImageInputsChanged` function, using the candidate's base-image workflow path rules, reports `false` for the full tested-commit-to-candidate diff. The changed paths are 16 documentation files and `test/cli/debug-command.test.ts`.
- Maintainer-directed evidence reuse: the maintainer explicitly selected the existing E2E run. This brief retains the earlier job's actual SHA and does not claim that a candidate-attached check passed. No new image publication is required by the intervening changes.
- Workflow bookkeeping: the docs-merge E2E run [34850495175](https://github.com/NVIDIA/NemoClaw/actions/runs/34850495175) is pending with no jobs. It is not the selected full-run evidence.
- Policy distinction: the release skill ordinarily asks for a candidate-attached successful aggregate. This preparation follows the maintainer's instruction to use the selected earlier E2E evidence plus the verified unchanged image inputs.

## General E2E context

- User-selected run: full E2E 34842707308, attempt 2. The maintainer explicitly chose to use this existing E2E run.
- databaseId: 34842707308
- attempt: 2
- headSha: 49765a4305ef1f61575f73cf7cb504dbc48d45db
- createdAt: 2026-09-14T13:11:04Z
- startedAt: 2026-09-14T13:11:01Z
- Last updated: 2026-09-14T13:39:42Z
- status: in_progress
- conclusion: not yet concluded
- url: https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/attempts/2
- Inspected: 2026-09-14T14:09:48.844625+00:00; age since creation: 111 minutes.
- The tested commit differs from the candidate. The intervening commits are [#11707](https://github.com/NVIDIA/NemoClaw/pull/11707), a CLI test-stub change, and docs merge [#11538](https://github.com/NVIDIA/NemoClaw/pull/11538). No intervening production-source change was found.
- Release qualification: not yet present in the active attempt.
- Snapshot at confirmation preparation: 72 successful jobs, 15 failed jobs, 9 skipped jobs, 1 running job. The run remains unresolved.

### Failed jobs and exception scope

Comparison: [full run 34814790468](https://github.com/NVIDIA/NemoClaw/actions/runs/34814790468) at `71223c18ce67db3ef17757805105f59182e89715`. Ten of the failed jobs passed there; four also failed there. This establishes newly observed outcomes, not a controlled regression proof or a verified common cause.

- [Hermes E2E (docker)](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988232594): Gateway restart never became healthy; connection refused. Previously failing job; prior result failure. [Prior job](https://github.com/NVIDIA/NemoClaw/actions/runs/34814790468/job/103883589793).
- [Networking: OpenClaw answers through balanced egress (docker) / Brave and NVIDIA inference API keys](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988236785): OpenClaw weather assertion received no reply and exit 255; cause remains unclassified. Newly observed failure; prior result success. [Prior job](https://github.com/NVIDIA/NemoClaw/actions/runs/34814790468/job/103883592950).
- [Pi: qualifies managed runtime on Linux AMD64 (docker) / NVIDIA API key](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988236881): Pi stop/start reached ready smoke checks, then the inference invocation returned HTTP 503. Newly observed failure; prior result success. [Prior job](https://github.com/NVIDIA/NemoClaw/actions/runs/34814790468/job/103883592406).
- [Upgrade: preserves a v0.0.89 sandbox on x86-64 (docker) / GitHub read token](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988237332): Upgrade installer exited nonzero; bounded excerpt does not establish the underlying cause. Previously failing job; prior result failure. [Prior job](https://github.com/NVIDIA/NemoClaw/actions/runs/34814790468/job/103883591962).
- [Messaging: OpenClaw preserves channels across stop and start (docker) / NVIDIA inference API key](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988237383): Channel lifecycle failed during rebuild after stopping all channels; cause remains unclassified. Newly observed failure; prior result success. [Prior job](https://github.com/NVIDIA/NemoClaw/actions/runs/34814790468/job/103883594054).
- [Gateway: restores the guard chain after recreation (docker) / NVIDIA inference API key](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988237481): Gateway guard recovery exhausted its polling bound. Previously failing job; prior result failure. [Prior job](https://github.com/NVIDIA/NemoClaw/actions/runs/34814790468/job/103883593146).
- [Rebuild: refreshes a stale Hermes base and restores state (docker) / NVIDIA inference API key](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988237968): Hermes cron restore recovery failed at publication after restoring its dashboard forward. Newly observed failure; prior result success. [Prior job](https://github.com/NVIDIA/NemoClaw/actions/runs/34814790468/job/103883593372).
- [Sandbox: preserves lifecycle and multi-sandbox operations (docker) / NVIDIA inference API key](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988238411): Multi-sandbox scenario failed during OpenClaw onboarding; two other tests passed. Newly observed failure; prior result success. [Prior job](https://github.com/NVIDIA/NemoClaw/actions/runs/34814790468/job/103883593250).
- [Rebuild: preserves Hermes state and recovers cron dispatch (docker) / NVIDIA inference API key](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988238525): Hermes cron restore recovery failed at publication after restoring its dashboard forward. Newly observed failure; prior result success. [Prior job](https://github.com/NVIDIA/NemoClaw/actions/runs/34814790468/job/103883593817).
- [Security: Hermes retains the required sandbox posture (docker) / NVIDIA inference API key](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988239314): Same Hermes gateway-restart connection-refused symptom as the general Hermes job; no security-escape conclusion follows. Previously failing job; prior result failure. [Prior job](https://github.com/NVIDIA/NemoClaw/actions/runs/34814790468/job/103883593749).
- [Performance: bounds hosted inference turns for OpenClaw and Hermes (docker) / NVIDIA inference API key](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988239520): Hermes preparation failed at publication/forward restoration before the intended latency measurement; not proof of a latency regression. Newly observed failure; prior result success. [Prior job](https://github.com/NVIDIA/NemoClaw/actions/runs/34814790468/job/103883593366).
- [Onboarding: Hermes reuses its sandbox and forwards (docker) / no provider credential](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988241010): Hermes second-sandbox agent setup failed after selecting a different dashboard port. Newly observed failure; prior result success. [Prior job](https://github.com/NVIDIA/NemoClaw/actions/runs/34814790468/job/103883592647).
- [Networking: OpenClaw reaches a public reference through open egress (docker) / Brave and NVIDIA inference API keys](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103989531507): OpenClaw onboarding exited nonzero before the public-reference assertion. Newly observed failure; prior result success. [Prior job](https://github.com/NVIDIA/NemoClaw/actions/runs/34814790468/job/103884343775).
- [Networking: Hermes reaches a public reference through open egress (docker) / Brave and NVIDIA inference API keys](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103990388647): The selected live E2E invocation ran no tests; coverage failure, not a demonstrated egress denial. Newly observed failure; prior result success. [Prior job](https://github.com/NVIDIA/NemoClaw/actions/runs/34814790468/job/103884418970).

- [Protected managed-image GPU and local inference](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103996496739): failed during OpenShell gateway startup in the GPU-backed Ollama qualification phase. Completed at 2026-09-14T14:05:33Z. Newly completed failure; no additional comparison proof is claimed.

### Other non-successful jobs

- [Exact staging Brev Launchable](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988233163): in_progress/not concluded; started 2026-09-14T13:13:29Z; completed not completed.
- [Exact staging Brev Launchable identity](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988233291): completed/skipped; started 2026-09-14T13:13:25Z; completed 2026-09-14T13:13:25Z.
- [native-runtime-qualification-producer-plan](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988233356): completed/skipped; started 2026-09-14T13:13:25Z; completed 2026-09-14T13:13:25Z.
- [openshell-dev-artifact](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988234297): completed/skipped; started 2026-09-14T13:13:25Z; completed 2026-09-14T13:13:25Z.
- [jetson-nvmap-gpu](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988234424): completed/skipped; started 2026-09-14T13:13:25Z; completed 2026-09-14T13:13:25Z.
- [external-gateway-health](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988234636): completed/skipped; started 2026-09-14T13:13:26Z; completed 2026-09-14T13:13:25Z.
- [MCP bridge dev (${{ matrix.agent }}, ${{ matrix.runtime_provider }})](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988239680): completed/skipped; started 2026-09-14T13:13:26Z; completed 2026-09-14T13:13:26Z.
- [Build pinned native Podman toolchain / ${{ matrix.architecture }}](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988239963): completed/skipped; started 2026-09-14T13:13:26Z; completed 2026-09-14T13:13:26Z.
- [matrix.jobName](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988248453): completed/skipped; started 2026-09-14T13:13:28Z; completed 2026-09-14T13:13:27Z.
- [Aggregate native runtime qualification evidence](https://github.com/NVIDIA/NemoClaw/actions/runs/34842707308/job/103988255556): completed/skipped; started 2026-09-14T13:13:29Z; completed 2026-09-14T13:13:29Z.

Staging Launchable cleanup is not yet verified because the job is still running. The release does not claim successful Launchable qualification or cleanup. Its final result and cleanup remain operational follow-up.

## Release carry-forward context

- Open target PRs: 31.
- Open target issues: 33.
- These unmerged items are outside the planned candidate. Automatic post-tag housekeeping owns carry-forward; no labels were changed.

- [PR #11725](https://github.com/NVIDIA/NemoClaw/pull/11725): fix(gateway): stop offering an inactive service as the stop path
- [PR #11693](https://github.com/NVIDIA/NemoClaw/pull/11693): fix(ci): scope SDK package lookup to the PR commit
- [PR #11617](https://github.com/NVIDIA/NemoClaw/pull/11617): fix(inference): admit OEM-chassis RTX Spark N1x to Ollama model ranking
- [PR #11616](https://github.com/NVIDIA/NemoClaw/pull/11616): fix(onboard): detect WSL Docker credential isolation from the caller's Docker selection
- [PR #11614](https://github.com/NVIDIA/NemoClaw/pull/11614): fix(sandbox): await deleted sandbox and report skipped gateway cleanup
- [PR #11612](https://github.com/NVIDIA/NemoClaw/pull/11612): fix(onboard): preserve legacy sandbox upgrade identity
- [PR #11582](https://github.com/NVIDIA/NemoClaw/pull/11582): fix(rebuild): reject expired inference credentials
- [PR #11530](https://github.com/NVIDIA/NemoClaw/pull/11530): fix(inference): adopt an interrupted managed vLLM container holding the serving port
- [PR #11528](https://github.com/NVIDIA/NemoClaw/pull/11528): fix(vllm): retire the last sandbox's managed vLLM container and probe its recorded port
- [PR #11405](https://github.com/NVIDIA/NemoClaw/pull/11405): fix(mcp): honor the recorded trusted private host in status discovery
- [PR #11389](https://github.com/NVIDIA/NemoClaw/pull/11389): fix(sandbox): isolate interrupted rebuild sessions
- [PR #11181](https://github.com/NVIDIA/NemoClaw/pull/11181): fix(snapshot): remove host Python sanitizer dependency
- [PR #11080](https://github.com/NVIDIA/NemoClaw/pull/11080): chore(deps): standardize Node 24 and npm 12
- [PR #10933](https://github.com/NVIDIA/NemoClaw/pull/10933): fix(cli): report the dashboard bind a sandbox was created with
- [PR #10860](https://github.com/NVIDIA/NemoClaw/pull/10860): fix(messaging): verify Telegram runtime credential
- [PR #10799](https://github.com/NVIDIA/NemoClaw/pull/10799): feat(install): add no-WSL Windows candidate installer
- [PR #10747](https://github.com/NVIDIA/NemoClaw/pull/10747): fix(cli): name the gateway that owns an unseen sandbox
- [PR #10733](https://github.com/NVIDIA/NemoClaw/pull/10733): fix(onboard): measure the host reuse window from collection completion
- [PR #10658](https://github.com/NVIDIA/NemoClaw/pull/10658): fix(uninstall): delete the shared Hugging Face cache NemoClaw created
- [PR #10629](https://github.com/NVIDIA/NemoClaw/pull/10629): fix(installer): normalize gateway service paths (#10541)
- [PR #10627](https://github.com/NVIDIA/NemoClaw/pull/10627): fix(ci): recognize current merge-gate jobs
- [PR #10498](https://github.com/NVIDIA/NemoClaw/pull/10498): fix(onboard): recover portable provider state
- [PR #10489](https://github.com/NVIDIA/NemoClaw/pull/10489): fix(messaging): preflight Teams webhook port
- [PR #10487](https://github.com/NVIDIA/NemoClaw/pull/10487): fix(policy): restrict raw GitHub methods
- [PR #10384](https://github.com/NVIDIA/NemoClaw/pull/10384): fix(onboard): account legacy credential aliases and stop deleting unread files
- [PR #10355](https://github.com/NVIDIA/NemoClaw/pull/10355): fix(pi): close release contract gaps
- [PR #10341](https://github.com/NVIDIA/NemoClaw/pull/10341): feat(installer): generalise deferred onboarding
- [PR #10245](https://github.com/NVIDIA/NemoClaw/pull/10245): fix(inference): forward ollama proxy bind-probe skip override to spawn env
- [PR #10200](https://github.com/NVIDIA/NemoClaw/pull/10200): fix(cli): redact list inventory rows like status rows
- [PR #10196](https://github.com/NVIDIA/NemoClaw/pull/10196): fix(debug): stage the debug bundle at an unpredictable owner-only path
- [PR #8910](https://github.com/NVIDIA/NemoClaw/pull/8910): feat(onboard): preserve Jetson GPU device groups

## General E2E decision

- Requested preparation: draft exceptions for failures that appear new while the run finishes.
- Maintainer choice: Proceed using the selected full E2E run and the approved exception rationale. After preparation, the maintainer requested “Ok, let’s tag.” The confirmation snapshot above includes the newly failed GPU job and still-running Launchable job.
- Approved rationale: Proceed with v0.0.124 to ship the documented improvements while investigating the listed failures separately; accept the earlier tested commit because subsequent changes affect only tests and documentation.
- Approval: the maintainer answered “yes” to this rationale in the release preparation task.
- Follow-up: record the final selected-run outcome and verify staging Brev Launchable cleanup when available. This release decision accepts that job as unresolved, not successful.

Exceptions: Proceed with v0.0.124 to ship the documented improvements while investigating the listed failures separately; accept the earlier tested commit because subsequent changes affect only tests and documentation. The selected full E2E run remains unresolved at this preparation snapshot, with 15 failed jobs and staging Brev Launchable still running; its cleanup is not yet verified. Existing successful image evidence is reused by maintainer direction; unchanged image inputs were verified.
Assets 2
Loading