feat(openshell): upgrade managed runtime to v0.0.99 - #8523
Conversation
Signed-off-by: Apurv Kumaria <akumaria@nvidia.com>
|
Note Reviews pausedIt looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the Use the following commands to manage reviews:
Use the checkboxes below for quick actions:
📝 WalkthroughWalkthroughThe PR upgrades NemoClaw’s OpenShell integration from 0.0.85 to 0.0.99. It updates pinned artifacts, credential manifests, compatibility rules, inference parsing, sandbox-name limits, E2E provenance, fixtures, and migration evidence. ChangesOpenShell release identity and artifact pins
Credential boundary manifest
Gateway and MCP compatibility contracts
Migration review and acceptance evidence
Estimated code review effort: 3 (Moderate) | ~30 minutes Possibly related issues
Possibly related PRs
Suggested labels: Suggested reviewers: 🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✨ Finishing Touches 💡 1📝 Generate docstrings 💡
🧪 Generate unit tests (beta)
Comment |
|
🌿 Preview your docs: https://nvidia-preview-pr-8523.docs.buildwithfern.com/nemoclaw |
PR Review Advisor — InformationalAdvisor assessment: Informational / low confidence Model lanes
Second-opinion terminology and E2E selections are advisory. Live E2E does not run automatically for pull requests. E2E guidanceAdvisory only. A maintainer can dispatch the default E2E suite against this exact revision. Recommended E2E: This automated review informs maintainers. Warnings and suggestions do not require a response. A maintainer decides whether to merge. |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@test/openshell-0.0.99-migration-review.test.ts`:
- Around line 38-41: The migration review test currently counts commit-like
tokens and checks string presence instead of validating evidence structure.
Replace the commitLedger token-count assertion with parsing of each range row,
asserting declared counts and unique commit IDs, and validate every archive,
extracted-binary, supervisor, source-commit, and manifest identity against
independent expected values tied to the 0.0.99 manifest. Update the runtime-gate
checks to assert observable behavior and artifact associations rather than
source-text occurrence.
🪄 Autofix
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: 1c11c0a1-b81b-4f83-8022-93cd63bfa75d
📒 Files selected for processing (51)
.github/workflows/e2e.yamlagents/hermes/Dockerfileagents/hermes/mcp-config-transaction.pydocs/deployment/set-up-mcp-bridge.mdxdocs/manage-sandboxes/add-mcp-server.mdxdocs/reference/commands.mdxdocs/reference/configure-runtime-identity.mdxdocs/reference/troubleshoot-mcp-servers.mdxdocs/reference/troubleshooting.mdxdocs/security/best-practices.mdxdocs/security/openshell-0.0.99-migration-review.mdnemoclaw-blueprint/blueprint.yamlscripts/brev-launchable-ci-cpu.shscripts/install-openshell.shscripts/update-hermes-agent.shsrc/lib/actions/sandbox/mcp-bridge-input-validation.test.tssrc/lib/actions/sandbox/mcp-bridge-validation.tssrc/lib/actions/sandbox/openshell-child-visible-credentials.v0.0.99.jsonsrc/lib/onboard/docker-driver-gateway-runtime.test.tssrc/lib/onboard/docker-driver-gateway-runtime.tssrc/lib/onboard/docker-gpu-patch-types.tssrc/lib/onboard/openshell-feature-gate.test.tssrc/lib/onboard/openshell-feature-gate.tssrc/lib/onboard/openshell-install.test.tssrc/lib/onboard/openshell-install.tssrc/lib/onboard/openshell-version.tstest/brev-launchable-ci-cpu-checksum.test.tstest/candidate-compat.test.tstest/deepagents-mcp-legacy-lifecycle.test.tstest/e2e/live/network-policy.test.tstest/e2e/live/openclaw-plugin-runtime-exdev.test.tstest/e2e/live/openshell-gateway-auth-source-contract-helpers.tstest/e2e/live/openshell-gateway-auth-source-contract.test.tstest/e2e/live/openshell-gateway-upgrade.test.tstest/fixtures/openshell-v0.0.99test/hermes-doctor-config-hash.test.tstest/hermes-final-image-layout.test.tstest/hermes-mcp-credential-boundary-manifest.test.tstest/install-openshell-version-check.test.tstest/mcp-add-crash-consistency.test.tstest/mcp-destroy-lifecycle.test.tstest/mcp-openshell-workflow.test.tstest/mcp-policy-key-ownership.test.tstest/mcp-restart-policy-order.test.tstest/openshell-0.0.99-migration-review.test.tstest/openshell-channel-workflow.test.tstest/pr-workflow-contract.test.tstest/sandbox-provisioning.test.tstest/sandbox-rlimit-hooks.test.tstest/update-hermes-agent-script.test.tstools/e2e/mcp-workflow-boundary.mts
Signed-off-by: Apurv Kumaria <akumaria@nvidia.com>
Signed-off-by: Apurv Kumaria <akumaria@nvidia.com>
|
@apurvvkumaria — I revalidated this review at head Product scope: PASS — #8497 authorizes the OpenShell v0.0.99 qualification and pin update. GitHub merge state: BLOCKED — GitHub reports the PR mergeable, but it is behind the current base and required Blocking checklist
I am holding approval until these items and the required GitHub gate are clear. |
Signed-off-by: Apurv Kumaria <akumaria@nvidia.com>
…e' into codex/8497-openshell-v099-upgrade
Signed-off-by: Apurv Kumaria <akumaria@nvidia.com>
…e' into codex/8497-openshell-v099-upgrade
|
Addressed all blockers from the maintainer review in verified commit
Fresh GitHub CI and exact all-agent managed startup are running on |
Signed-off-by: Apurv Kumaria <akumaria@nvidia.com>
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@src/lib/actions/sandbox/mcp-bridge-url-validation.ts`:
- Around line 224-227: Update the DNS-pinning validation and its explanatory
comment near the OpenShell connection flow to require
proxy_connect_by_hostname=false; reject or fail validation when the setting is
enabled, so the allowed_ips guarantee only applies when OpenShell connects using
its pinned SocketAddr list.
In `@test/runner.test.ts`:
- Around line 368-370: Extend the validateName boundary tests around the
existing 19-character case to assert that the first invalid length, 20
characters, throws the same “sandbox name too long (max 19 chars)” error. Keep
the existing valid 19-character and oversized-input assertions unchanged.
🪄 Autofix
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: 319fa606-10d0-4b11-a7cc-d8a16414b59e
📒 Files selected for processing (20)
agents/hermes/mcp-config-transaction.pyci/source-shape-test-budget.jsondocs/deployment/set-up-mcp-bridge.mdxdocs/reference/commands.mdxdocs/reference/troubleshooting.mdxdocs/security/openshell-0.0.99-migration-review.mdnemoclaw/src/shared/openshell-policy-boundary.ctsnemoclaw/src/shared/sandbox-name.ctssrc/lib/actions/sandbox/mcp-bridge-url-validation.tssrc/lib/deploy/index.test.tssrc/lib/onboard/forward-start.tstest/e2e/support/mcp-bridge-sandbox.test.tstest/installer-hash-check.test.tstest/onboard-sandbox-name.test.tstest/openshell-0.0.85-migration-review.test.tstest/openshell-0.0.99-migration-review.test.tstest/repro-5978-policy-denial-hint.test.tstest/runner.test.tstest/update-hermes-agent-script.test.tstools/e2e/cli-artifact-workflow-boundary.mts
🚧 Files skipped from review as they are similar to previous changes (5)
- docs/security/openshell-0.0.99-migration-review.md
- agents/hermes/mcp-config-transaction.py
- docs/reference/commands.mdx
- test/update-hermes-agent-script.test.ts
- docs/reference/troubleshooting.mdx
Signed-off-by: Apurv Kumaria <akumaria@nvidia.com>
|
Follow-up exact-activation blocker addressed in verified commits |
Signed-off-by: Apurv Kumaria <akumaria@nvidia.com>
Signed-off-by: Apurv Kumaria <akumaria@nvidia.com>
Signed-off-by: Apurv Kumaria <akumaria@nvidia.com>
|
@apurvvkumaria — one current-head blocker remains on
The exact all-agent activation, documentation receipt, CodeRabbit, and published PR Review Advisor result are otherwise clear on this head. |
Signed-off-by: Apurv Kumaria <akumaria@nvidia.com>
|
Additional #8497 activation evidence from unrelated PR #8560: Managed Images run 31186757641, job 92896129767, used OpenShell v0.0.85. OpenClaw activation/recovery/cleanup passed, then Hermes sandbox creation ended after 77 seconds with terminal_failure_phase before runtime exercise. #8523's current v0.0.99 all-agent activation passed in run 31186050945, job 92893723945. This maps to #8497's accepted v0.0.85 managed-image activation workstream and remains owned here; no duplicate E2E run was dispatched. |
## Summary Base-image publication can fail immediately after pushing a candidate manifest when GHCR's token endpoint transiently returns `403 Forbidden` to the first manifest read. This change gives registry reads five bounded attempts while preserving the existing digest, platform, and provenance verification. The failure was observed in main E2E run 31184501275, job 92888376997, propagated from Base Images run 31184501811, job 92886728421 (`Build and push Hermes base image`). ## Changes - Route base-image manifest reads through one fixed-policy retry helper. - Preserve the final Docker exit status after five unsuccessful attempts. - Keep all existing source-digest, platform, candidate-manifest, and publication checks unchanged. - Add focused helper, workflow-shape, and opaque-input trigger coverage. ## Type of Change - [x] Code change (feature, bug fix, or refactor) - [ ] Code change with doc updates - [ ] Doc only (prose changes, no code sample modifications) - [ ] Doc only (includes code sample changes) ## Quality Gates - [x] Tests added or updated for changed behavior - [ ] Existing tests cover changed behavior — justification: - [ ] Tests not applicable — justification: - [ ] Docs updated for user-facing behavior changes - [x] Docs not applicable — justification: This changes only internal GitHub Actions registry-read reliability; no user-facing command, configuration, default, or supported product behavior changes. - [x] Sensitive paths changed (security, policy, credentials, preflight, onboarding, inference, runner, sandbox, or messaging) - [x] Sensitive-path review completed or maintainer-approved waiver recorded — reviewer/approval link/justification: Maintainer nine-category review passed. The retry is bounded, arguments remain quoted, terminal failures propagate, temporary stderr is removed, credential handling is unchanged, and digest/platform/provenance verification remains fail-closed. - [x] Non-success, skipped, or missing CI check accepted by maintainer — check name, approval link, and follow-up issue: `PR exact all-agent managed runtime activation` failed in [run 31186757641, job 92896129767](https://github.com/NVIDIA/NemoClaw/actions/runs/31186757641/job/92896129767) on the existing OpenShell v0.0.85 Hermes sandbox-creation path. Accepted issue #8497 owns that v0.0.85 activation failure, and PR #8523's v0.0.99 activation passed in [run 31186050945, job 92893723945](https://github.com/NVIDIA/NemoClaw/actions/runs/31186050945/job/92893723945). The classification is recorded in [#8560](#8560 (comment)) and routed to [#8523](#8523 (comment)). ## Documentation Writer Review - [x] Documentation writer subagent reviewed the completed changes - Result: `no-docs-needed` - Evidence: The change only adds bounded retries to internal GitHub Actions base-image registry reads. It does not change a user-facing command, configuration, default, or supported product behavior. - Agent: Codex Desktop <!-- docs-review-head-sha: 069dbd4 --> <!-- docs-review-agents-blob-sha: 12ad395 --> ## DGX Station Hardware Evidence - [ ] Tested on DGX Station - Tested commit: Not applicable. - Station profile/scenario: Not applicable. - Result: Not applicable. - Supporting evidence: `scripts/prepare-dgx-station-host.sh` is unchanged. ## Verification - [x] PR description includes a `Signed-off-by:` line and every commit appears as `Verified` in GitHub - [x] Normal `pre-commit`, `commit-msg`, and `pre-push` hooks passed, or `npm run validate:pr` passed after refreshing `origin/main` when hooks were skipped or unavailable - [x] Targeted behavior tests pass for the current change set, or tests are marked not applicable above — `npx vitest run --project integration test/retry-docker-imagetools-inspect.test.ts test/dcode-base-image-workflow.test.ts test/managed-image-publication-workflow.test.ts test/vitest-watch-triggers.test.ts` (4 files, 30 tests passed) - [ ] Applicable broad gate passed — `npm test` for broad runtime/test-harness changes; `npm run check` for repo-wide validation/coverage changes — command/result: Not applicable to this narrow internal workflow retry. Normal repository checks passed. - [x] Quality Gates section completed with required justifications or waivers - [x] No secrets, API keys, or credentials committed - [ ] `npm run docs` builds without warnings (doc changes only) - [ ] Doc pages follow the [style guide](https://github.com/NVIDIA/NemoClaw/blob/main/docs/CONTRIBUTING.md) (doc changes only) - [ ] New doc pages include SPDX header and frontmatter (new pages only) The existing `validate-managed-base-index` suite requires Bash associative arrays. This macOS host has only Bash 3.2, so that suite could not execute locally; required Linux CI remains authoritative for that integration path. --- Signed-off-by: Senthil Ravichandran <senthilr@nvidia.com> Signed-off-by: Senthil Ravichandran <senthilr@nvidia.com>
|
Addressed in signed, GitHub-verified commit |
# Conflicts: # tools/e2e/cli-artifact-workflow-boundary.mts
|
The release-evidence blocker remains addressed in signed, GitHub-verified commit The two newly merged upstream changes have no source/test overlap with the OpenShell upgrade; their shared documentation edits combined cleanly. On the final tree, 114 upstream GPU tests, 23 OpenShell migration/release-evidence tests, and 62 llama.cpp lifecycle tests pass. Fresh required CI, automated review, and exact all-agent managed activation are now running on the updated revision. |
senthilr-nv
left a comment
senthilr-nv
left a comment
There was a problem hiding this comment.
Reviewed 18ac71c45 against current base aae23eb4f.
Required checks and exact all-agent managed activation pass. The corrected release-evidence contract passes 19/19 locally; the exact-head documentation-writer receipt is current. CodeRabbit has no unresolved finding, and the informational PR Review Advisor result reports zero findings.
Product scope is approved under #8497.
Signed-off-by: Apurv Kumaria <akumaria@nvidia.com>
Signed-off-by: Apurv Kumaria <akumaria@nvidia.com>
<!-- markdownlint-disable MD041 --> ## Summary Completes the remaining OpenShell `0.0.99` compatibility work after #8499 and #8523. Existing `0.0.85` sandboxes can now survive the gateway upgrade without losing the required supervisor workdir argument, and Docker/Podman lifecycle authority matches the exact `0.0.99` identity and status contracts. ## Related Issue Closes #8497 ## Changes - Preserve only the reviewed OpenShell supervisor command (`--workdir /sandbox`) when cloning a managed workload, while retaining the empty-command form needed by `0.0.85` migration and rejecting other supervisor arguments before mutation. Focused clone/startup tests protect both accepted forms and the fail-closed path. - Bind Docker and Podman lifecycle mutations to the exact `0.0.99` default-workspace identity: workspace-qualified names, exact labels, and full immutable container IDs. Ownership and lifecycle tests reject foreign workspaces, ambiguous containers, and legacy Podman identities. - Pin the portable Podman gateway to the prepared rootless socket and add a credential-free live proof using OpenShell `0.0.99` and Podman `5.7`. The lane disables Docker, checks authenticated gateway health, activates all registered agent identities, and exercises the production portable lifecycle binding. - Scope gateway status probes to the requested gateway and recognize the `0.0.99` connection-error form that omits the gateway name. OpenShell does not yet provide a structured status error kind, so the fallback remains limited to an explicitly scoped gateway and is covered by exact-output fixtures. - Update the migration/security review and pinned-version references with the completed compatibility contracts and remaining live-CI qualification gate. ## Type of Change - [ ] Code change (feature, bug fix, or refactor) - [x] Code change with doc updates - [ ] Doc only (prose changes, no code sample modifications) - [ ] Doc only (includes code sample changes) ## Quality Gates - [x] Tests added or updated for changed behavior - [ ] Existing tests cover changed behavior — justification: - [ ] Tests not applicable — justification: - [x] Docs updated for user-facing behavior changes - [ ] Docs not applicable — justification: - [x] Sensitive paths changed (security, policy, credentials, preflight, onboarding, inference, runner, sandbox, or messaging) - [x] Sensitive-path review completed or maintainer-approved waiver recorded — reviewer/approval link/justification: Independent Codex Desktop review found no blockers after checking the exact OpenShell `0.0.99` source contracts, ownership boundaries, rootless socket binding, fail-closed behavior, and live-workflow cleanup. - [ ] Non-success, skipped, or missing CI check accepted by maintainer — check name, approval link, and follow-up issue: ## Documentation Writer Review - [x] Documentation writer subagent reviewed the completed changes - Result: `docs-updated` - Evidence: `docs/security/openshell-0.0.99-migration-review.md`, `docs/security/openclaw-2026.7.1-dependency-review.md`, and `tools/pr-review-advisor/README.md` - Agent: Codex Desktop <!-- docs-review-head-sha: 3b3f3e2 --> <!-- docs-review-agents-blob-sha: 12ad395 --> ## DGX Station Hardware Evidence - [ ] Tested on DGX Station - Tested commit: Not applicable; DGX Station preparation did not change. - Station profile/scenario: - Result: - Supporting evidence: ## Verification - [x] PR description includes a `Signed-off-by:` line and every commit appears as `Verified` in GitHub - [x] Normal `pre-commit`, `commit-msg`, and `pre-push` hooks passed, or `npm run validate:pr` passed after refreshing `origin/main` when hooks were skipped or unavailable - [x] Targeted behavior tests pass for the current change set, or tests are marked not applicable above — 308 focused compatibility tests passed before the final E2E correction; post-merge ownership/session coverage passed 81/81, Podman support coverage passed 9/9, the affected destroy integration coverage passed 4/4, semantic E2E coverage passed for 122 tests across 79 files, and CLI type-check plus architecture checks passed. - [ ] Applicable broad gate passed — GitHub CI will run the complete matrix. A local repository-wide fast run was not claimed because unrelated timing-sensitive tests flaked under local load; all affected-area suites pass. - [x] Quality Gates section completed with required justifications or waivers - [x] No secrets, API keys, or credentials committed - [ ] `npm run docs` builds without warnings (doc changes only) — completed with 0 errors and 2 pre-existing Fern warnings. - [x] Doc pages follow the [style guide](https://github.com/NVIDIA/NemoClaw/blob/main/docs/CONTRIBUTING.md) (doc changes only) - [ ] New doc pages include SPDX header and frontmatter (new pages only) --- Signed-off-by: Apurv Kumaria <akumaria@nvidia.com> <!-- This is an auto-generated comment: release notes by coderabbit.ai --> ## Summary by CodeRabbit * **New Features** * Added rootless Podman socket support for portable profiles. * Improved workspace-aware sandbox discovery, container ownership, and SSH host routing. * Added gateway-specific status checks and deletion-convergence validation. * **Bug Fixes** * Prevented ambiguous, malformed, or mismatched containers from being reused or modified. * Preserved container commands during Docker-based cloning while rejecting unsupported configurations. * Improved handling of interrupted command results and inference route detection. * **Testing** * Expanded end-to-end coverage for Podman lifecycle, networking, cleanup, diagnostics, and gateway startup. * **Documentation** * Updated compatibility and migration guidance for OpenShell 0.0.99. <!-- end of auto-generated comment: release notes by coderabbit.ai --> --------- Signed-off-by: Apurv Kumaria <akumaria@nvidia.com> Signed-off-by: Prekshi Vyas <prekshiv@nvidia.com> Co-authored-by: Carlos Villela <cvillela@nvidia.com> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: Prekshi Vyas <prekshiv@nvidia.com>
4 participants
Summary
Upgrade NemoClaw's consumed OpenShell runtime from v0.0.85 to v0.0.99 after #8499 pinned the exact v0.0.99 release manifests. This PR pins the corresponding CLI, gateway, sandbox, supervisor, and child-visible credential identities across install, blueprint, managed-agent, Brev, and E2E paths. It also resolves v0.0.99 policy-activation incompatibilities for OpenClaw npm, Homebrew's GitHub endpoints, and Outlook's shared Microsoft endpoints.
Related Issue
Part of #8497.
Changes
Type of Change
Quality Gates
docs/security/openshell-0.0.99-migration-review.mdandsrc/lib/actions/sandbox/openshell-child-visible-credentials.v0.0.99.json; repository-owned maintainer review routing remains authoritative.Documentation Writer Review
docs-updateddocs/deployment/set-up-mcp-bridge.mdx,docs/manage-sandboxes/add-mcp-server.mdx,docs/manage-sandboxes/update-sandboxes.mdx,docs/network-policy/create-custom-policy-presets.mdx,docs/reference/architecture.mdx,docs/reference/commands.mdx,docs/reference/configure-runtime-identity.mdx,docs/reference/network-policies.mdx,docs/reference/troubleshoot-mcp-servers.mdx,docs/reference/troubleshooting.mdx,docs/security/best-practices.mdx,docs/security/openshell-0.0.72-compatibility-review.mdx, anddocs/security/openshell-0.0.99-migration-review.mdDGX Station Hardware Evidence
Verification
Signed-off-by:line and every commit appears asVerifiedin GitHubpre-commit,commit-msg, andpre-pushhooks passed, ornpm run validate:prpassed after refreshingorigin/mainwhen hooks were skipped or unavailable18ac71c45passed 35 release/maintainer tests and 144 workflow-boundary tests. On final revision6d5e9568a, 114 upstream GPU tests, 23 OpenShell migration/release-evidence tests, and 62 llama.cpp lifecycle tests passed;npm run typecheck:cliandnpm run docspassed, normal merge hooks passed, and pre-push type/version gates passed.npm run validate:prpassed on parent revision509a5bc60.npm testfor broad runtime/test-harness changes;npm run checkfor repo-wide validation/coverage changes — command/result:npm testwas attempted locally but the heavily loaded host produced widespread unrelated loopback/timing failures across untouched suites; changed-area suites pass in isolation and fresh PR CI is required.npm run docsbuilds without warnings (doc changes only) — the final revision build passed with 0 errors and 2 existing Fern warnings, so this remains unchecked.Signed-off-by: Apurv Kumaria akumaria@nvidia.com
Summary by CodeRabbit
New Features
Bug Fixes
Documentation
Tests