ci: auto-update release notes on push to main by ericksoa · Pull Request #1 · NVIDIA/NemoClaw

ericksoa · 2026-03-16T18:25:53Z

Migrated from NVIDIA/openshell-openclaw-plugin#31

Summary

Adds a GitHub Actions workflow that auto-updates docs/about/release-notes.md on every push to main
Categorizes commits into Features / Fixes / Other based on commit message prefixes (case-insensitive)
Filters out noise: reverts, chores, docs-only, test-only, style, and low-signal commits
Scopes to only new commits since last run (uses [release-notes] marker in commit message)
Self-loop prevention: skips if the triggering commit is from this workflow
Commits back with -s (DCO signed-off)

How it works

On push to main, collects non-merge commits since last tag or last [release-notes] commit
Categorizes by prefix: add/feat → Features, fix → Fixes, everything else → Other
Inserts a dated block under the ## 0.1.0 Unreleased heading
Commits and pushes the updated file

Test plan

Merge a commit to main and verify the workflow runs
Check that docs/about/release-notes.md gets a new dated section
Verify the workflow doesn't trigger itself (no infinite loop)
Verify commit messages are correctly categorized

…to main Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Signed-off-by: Aaron Erickson <aerickson@nvidia.com>

Add a preflight check that catches the #1 onboarding blocker on Ubuntu 24.04, DGX Spark, and WSL2. When cgroup v2 is active but Docker's daemon.json lacks "default-cgroupns-mode": "host", onboarding now fails fast with a clear error and fix instructions instead of failing late at gateway startup with a cryptic kubelet error. Closes #16

wscurran · 2026-03-20T23:17:23Z

Thanks for setting up a GitHub Actions workflow that automatically updates the release notes on pushes to the main branch, which could help keep the documentation up to date and reduce manual effort.

cv · 2026-03-21T19:27:31Z

Hi @ericksoa! Thanks for putting this together — auto-updating release notes is something we've been wanting. Since this was opened, the repo has seen a lot of activity: we've added CI checks, new features, and restructured a few things. Would you mind rebasing onto the latest main when you get a chance? That way we can give it a proper review with everything up to date. Appreciate it!

ericksoa · 2026-03-22T16:30:51Z

Superseded — release notes page now points to GitHub native releases/commits/PRs.

- Default model: nvidia/nemotron-3-super-120b-a12b (March 2026 release, 12B active / 120B total MoE, 5x throughput) - Replace meta/llama-3.3-70b-instruct with Nemotron 3 Super in the nvidia provider catalog (Dockerfile sed patch at build time) - All NVIDIA models in provider: Nemotron 70B, Mistral NeMo Minitron, Nemotron 3 Super — no Meta/Llama - Blueprint default profile updated to nemotron-3-super-120b-a12b - Tracked as tech debt: NVIDIA#1 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

…IA#62) Add a preflight check that catches the NVIDIA#1 onboarding blocker on Ubuntu 24.04, DGX Spark, and WSL2. When cgroup v2 is active but Docker's daemon.json lacks "default-cgroupns-mode": "host", onboarding now fails fast with a clear error and fix instructions instead of failing late at gateway startup with a cryptic kubelet error. Closes NVIDIA#16

…docs Add two behavioral tests that directly validate cv's blocker NVIDIA#1 fix: - Healthy gateway is preserved (no destroy/forward-stop) on rerun - Stale vs healthy vs active-unnamed states trigger correct cleanup Also add setup-jetson entry to docs/reference/commands.md. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

… (#1305) ## Summary Fixes the four issues reported in #1114 — EACCES permission errors and missing gateway token when running inside the NemoClaw sandbox. ### Issue mapping | # | Reported error | Fix | |---|----------------|-----| | 1 | `EACCES: open '/sandbox/.openclaw/openclaw.json.*.tmp'` | `install_configure_guard` — intercepts `openclaw configure` with a clear error and directs users to `nemoclaw onboard --resume` on the host | | 2 | Same as #1 (different PID) | Same fix | | 3 | `EACCES: mkdir '/sandbox/.openclaw/credentials'` | Already resolved on main via #1519 (credentials symlink to `.openclaw-data/`) | | 4 | No WhatsApp QR code | Consequence of #3, also resolved by #1519 | ### Root cause (issues 1 & 2) OpenClaw's `configure` command performs atomic writes — it creates a temp file (`openclaw.json.PID.UUID.tmp`) in the same directory as the config. Since `/sandbox/.openclaw/` is Landlock read-only at the kernel level, file creation is rejected with EACCES. This is by design: the sandbox config is intentionally immutable at runtime. Rather than weakening Landlock (security regression), we intercept the command in the sandbox shell and guide users to the correct host-side workflow. ### Changes **1. `install_configure_guard()`** — Writes a shell function wrapper to `.bashrc`/`.profile` that intercepts `openclaw configure` and prints: ``` Error: 'openclaw configure' cannot modify config inside the sandbox. The sandbox config is read-only (Landlock enforced) for security. To change your configuration, exit the sandbox and run: nemoclaw onboard --resume This rebuilds the sandbox with your updated settings. ``` All other `openclaw` subcommands pass through to the real binary. **2. `export_gateway_token()`** — Reads `gateway.auth.token` from `openclaw.json` and exports it as `OPENCLAW_GATEWAY_TOKEN`, so interactive sessions (`openshell sandbox connect`) can authenticate with the gateway. Persists to `.bashrc`/`.profile` using idempotent marker blocks and cleans stale tokens on revocation. **3. `_read_gateway_token()` helper** — Shared Python snippet used by both `export_gateway_token` and `print_dashboard_urls` (deduplication, uses `with open()` context manager). All three are called in both root and non-root startup paths. ## Security properties preserved - `/sandbox/.openclaw` remains root-owned, Landlock read-only - `openclaw.json` remains chmod 444 (immutable) - No new attack surface — token is read-only from existing config - `command openclaw` bypass preserves all non-configure functionality Fixes #1114 Signed-off-by: Dongni Yang <dongniy@nvidia.com> Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com> Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> --------- Signed-off-by: Dongni Yang <dongniy@nvidia.com> Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>

…IA#1114) (NVIDIA#1305) ## Summary Fixes the four issues reported in NVIDIA#1114 — EACCES permission errors and missing gateway token when running inside the NemoClaw sandbox. ### Issue mapping | # | Reported error | Fix | |---|----------------|-----| | 1 | `EACCES: open '/sandbox/.openclaw/openclaw.json.*.tmp'` | `install_configure_guard` — intercepts `openclaw configure` with a clear error and directs users to `nemoclaw onboard --resume` on the host | | 2 | Same as NVIDIA#1 (different PID) | Same fix | | 3 | `EACCES: mkdir '/sandbox/.openclaw/credentials'` | Already resolved on main via NVIDIA#1519 (credentials symlink to `.openclaw-data/`) | | 4 | No WhatsApp QR code | Consequence of NVIDIA#3, also resolved by NVIDIA#1519 | ### Root cause (issues 1 & 2) OpenClaw's `configure` command performs atomic writes — it creates a temp file (`openclaw.json.PID.UUID.tmp`) in the same directory as the config. Since `/sandbox/.openclaw/` is Landlock read-only at the kernel level, file creation is rejected with EACCES. This is by design: the sandbox config is intentionally immutable at runtime. Rather than weakening Landlock (security regression), we intercept the command in the sandbox shell and guide users to the correct host-side workflow. ### Changes **1. `install_configure_guard()`** — Writes a shell function wrapper to `.bashrc`/`.profile` that intercepts `openclaw configure` and prints: ``` Error: 'openclaw configure' cannot modify config inside the sandbox. The sandbox config is read-only (Landlock enforced) for security. To change your configuration, exit the sandbox and run: nemoclaw onboard --resume This rebuilds the sandbox with your updated settings. ``` All other `openclaw` subcommands pass through to the real binary. **2. `export_gateway_token()`** — Reads `gateway.auth.token` from `openclaw.json` and exports it as `OPENCLAW_GATEWAY_TOKEN`, so interactive sessions (`openshell sandbox connect`) can authenticate with the gateway. Persists to `.bashrc`/`.profile` using idempotent marker blocks and cleans stale tokens on revocation. **3. `_read_gateway_token()` helper** — Shared Python snippet used by both `export_gateway_token` and `print_dashboard_urls` (deduplication, uses `with open()` context manager). All three are called in both root and non-root startup paths. ## Security properties preserved - `/sandbox/.openclaw` remains root-owned, Landlock read-only - `openclaw.json` remains chmod 444 (immutable) - No new attack surface — token is read-only from existing config - `command openclaw` bypass preserves all non-configure functionality Fixes NVIDIA#1114 Signed-off-by: Dongni Yang <dongniy@nvidia.com> Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com> Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> --------- Signed-off-by: Dongni Yang <dongniy@nvidia.com> Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>

- Guard runArgv/runArgvCapture against shell:true to prevent security bypass (finding #1) — throws if a caller attempts to re-enable shell interpretation. Added 2 tests. - Document the intentional bash -c exception in getOllamaWarmupCommand explaining why it's safe (finding NVIDIA#2). - Remove dead getOpenshellCommand() from policies.ts (finding NVIDIA#3). - Remove unused shellQuote import from nim.ts (finding NVIDIA#4). - Fix brittle indexOf assertion in onboard-readiness test (finding NVIDIA#5).

…xercised CodeRabbit correctly flagged that the original test swapped the lock file on stat NVIDIA#1 which caused isProcessAlive to see a live PID and exit early — unlinkIfInodeMatches was never called. Move the swap to after stat NVIDIA#1 returns so the stale-cleanup branch is actually reached and the inode comparison is tested. Signed-off-by: Prekshi Vyas <prekshiv@nvidia.com>

…exercised CodeRabbit correctly flagged that swapping on stat NVIDIA#1 caused readFileSync to see the live PID and exit via isProcessAlive — unlinkIfInodeMatches was never called. Move the swap to just before stat NVIDIA#2 (inside unlinkIfInodeMatches): stat NVIDIA#1 reads the original stale inode, readFileSync sees the dead PID, isProcessAlive returns false, stale-cleanup runs, and stat NVIDIA#2 sees the new inode and skips the unlink. Signed-off-by: Prekshi Vyas <prekshiv@nvidia.com>

…exercised CodeRabbit correctly flagged that swapping on stat NVIDIA#1 caused readFileSync to see the live PID and exit via isProcessAlive — unlinkIfInodeMatches was never called. Move the swap to just before stat NVIDIA#2 (inside unlinkIfInodeMatches): stat NVIDIA#1 reads the original stale inode, readFileSync sees the dead PID, isProcessAlive returns false, stale-cleanup runs, and stat NVIDIA#2 sees the new inode and skips the unlink. Use write-to-temp + rename instead of unlink + recreate to guarantee a different inode even on tmpfs/overlayfs which can reuse inodes. Signed-off-by: Prekshi Vyas <prekshiv@nvidia.com>

- Remove unused getForwardList() call from getActiveSandboxSessions — only pgrep/ps is needed for SSH session detection (warning #1) - Consolidate double-prompt in sandboxDestroy into single enriched confirmation prompt (warning #2) - Remove noisy cleanupGatewayAfterLastSandbox forward check that would always fire due to dashboard forward (warning #3) - Use word-boundary regex in parseSshProcesses to prevent false positives when sandbox names share prefixes (warning #4) - Export SessionClassification as named interface (suggestion #1) - Use cross-platform ps -axo instead of Linux-only pgrep -a for macOS compatibility (suggestion #2) - Add forwardCount to SessionClassification for future consumers - Add tests for word-boundary matching edge cases

…ded on main Drift from main — test-hermes-inference-switch.sh, test-openclaw-inference-switch.sh, and test-openshell-gateway-upgrade.sh were added to main after this branch forked. The convention lint correctly flagged them as orphans (Risk #1: new legacy scripts with no migration mapping). Added placeholder entries (empty scenario + assertions) to bring the map back in sync with the legacy surface area. Actual mappings land with their respective migration waves.

Resolve the two output threads in #3456 left after the core dead-loop fix landed via #3459 + #3434: Sub-bug #3 — `src/lib/onboard.ts` printed `nemoclaw <name> destroy --yes && nemoclaw onboard --gpu` with a literal `<name>` placeholder, and assumed at least one sandbox was registered. When the GPU-passthrough mismatch hit on the State B re-run path with an empty registry (the dead-loop case), the hint was not actionable. Replace with a registry-aware helper at `src/lib/onboard/gpu-recovery.ts` that renders the right shape: - empty registry → suggest `nemoclaw uninstall && nemoclaw onboard --gpu` - one sandbox → suggest destroy --yes --cleanup-gateway for that name - multiple sandboxes → list each, only the last gets --cleanup-gateway Sub-bug #4 — `src/lib/actions/uninstall/run-plan.ts` printed `Destroyed gateway 'nemoclaw' skipped` when the openshell destroy no-op'd (gateway already gone) — the "Destroyed … skipped" wording was self-contradictory. Extend `runOptional` with an `onSkip` option; route the gateway destroy to emit `Gateway 'nemoclaw' already removed or unreachable` on no-op. Tests: - `src/lib/onboard/gpu-recovery.test.ts` (6 tests): forbid literal `<name>` placeholder anywhere in the output; cover empty / single / multi-sandbox cases; defensive filter on whitespace names so a `nemoclaw destroy` rendering can never happen. - `src/lib/actions/uninstall/run-plan.test.ts`: assert the new "already removed or unreachable" wording and the absence of the "Destroyed gateway 'nemoclaw' skipped" string. The core dead loop itself (sub-bugs #1, #2 and State B GPU mismatch) is already addressed by #3459 + #3434 + #3483; #3456 will close once this lands. See the #3456 status comment for the full mapping. Refs #3456. Mirrors (and tightens) the approach in the closed PR #3464, which left the literal `<name>` placeholder in tests per CodeRabbit feedback that was never addressed. Signed-off-by: Charan Jagwani <charjags100@gmail.com>

…3520) > **Draft for visibility.** Issue-autopilot Stages 4-5 of #3456. Will mark ready once batch self-review + CI complete. ## Summary Closes the two remaining output threads in #3456 after the core dead-loop fix already landed on `main` (via #3459, #3434, #3483). Full sub-bug mapping in the [#3456 status comment](#3456 (comment)). - **Sub-bug #3** — `nemoclaw <name> destroy --yes` recovery hint replaced with a registry-aware helper. - **Sub-bug #4** — `Destroyed gateway 'nemoclaw' skipped` self-contradictory wording replaced with `Gateway 'nemoclaw' already removed or unreachable`. ## Acceptance criteria mapping | Sub-bug | Resolution | Evidence | |---|---|---| | #1 dead loop | Already fixed on main (#3459) | out of scope | | #2 firewall diagnostic | Already fixed on main (#3459) | out of scope | | **#3** literal `<name>` placeholder | **This PR** | `src/lib/onboard/gpu-recovery.ts` + `onboard.ts:10387-10405` | | **#4** misleading "skipped" wording | **This PR** | `src/lib/actions/uninstall/run-plan.ts:210-228, 407-414` | | #5 uninstall residuals | Already fixed on main (#3483) | out of scope | ## Behavior matrix `gpuPassthroughRecoveryLines(names)`: | Input | Suggestion | |---|---| | `null` / `[]` | `nemoclaw uninstall && nemoclaw onboard --gpu` | | one sandbox | `nemoclaw <name> destroy --yes --cleanup-gateway && nemoclaw onboard --gpu` | | many sandboxes | each `destroy --yes`, only the last gets `--cleanup-gateway` | ## Test plan ``` npm run typecheck:cli npx vitest run src/lib/onboard/gpu-recovery.test.ts src/lib/actions/uninstall/run-plan.test.ts ``` 22 tests pass (6 new + 16 existing). ## Notes for reviewers - This is the work [#3464 attempted](#3464); that PR was closed without merging after CodeRabbit asked for the `<name>` placeholder to be forbidden in tests via negative assertion. This PR adopts that refinement. - `runOptional` extension is backwards-compatible — existing callers without `onSkip` get the original wording. Closes #3456 once merged. --------- Signed-off-by: Charan Jagwani <charjags100@gmail.com> Co-authored-by: Charan Jagwani <charjags100@gmail.com> Co-authored-by: Carlos Villela <cvillela@nvidia.com>

Two streaming bugs that combined to hang every open-webui chat with the chad model for ~60s per turn until the frontend's read timeout: 1. The handler advertised `Connection: keep-alive` for the SSE response. Open-webui's frontend held the TCP connection open waiting for more chat-completion chunks (legitimate for normal HTTP, wrong for an SSE stream that's done after the [DONE] sentinel). Switched to `Connection: close` and explicit `self.close_connection = True` so BaseHTTPRequestHandler tears down the socket the moment the handler returns. 2. No flush between SSE events. Python's BufferedWriter on BaseHTTPRequestHandler buffers writes until the socket closes — which, combined with bug NVIDIA#1, meant the entire response landed at once *after* the connection should have already been done. Added explicit self.wfile.flush() calls after each event so chunks arrive incrementally. Symptom seen: open-webui showed a 2/2 swipe carousel where the actual chat reply briefly appeared and then got replaced; the "thinking bubble" animation kept playing after the response finished. The 2/2 was secondary — open-webui's auto-generated title (now routed to a separate fast NVIDIA endpoint via the task-model config) was racing with the chat reply because both streams were still "open" from open-webui's view. Verified live: streaming round-trip is now 12s (was hitting the 60s wall every time). Non-streaming path unchanged. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Run 26550936453 surfaced two more real bugs after Phase 6 wiring went live. Both fixed inside the spec's prescribed layers; nothing leaks into workflow YAML or clients. 1. dispatch-action.sh called e2e_context_init unconditionally before sourcing the install/onboard dispatcher. e2e_context_init opens context.env with `: > ctx`, which truncated the file the ScenarioRunner had just seeded. All runtime assertions then failed with 'e2e context: missing required key(s): E2E_SCENARIO ...'. Fix: dispatch-action.sh no longer calls e2e_context_init. The TS framework owns context.env initialization; workers may still extend it via e2e_context_set. 2. The legacy onboarding.preflight.passed assertion expects an onboard.log file at ${E2E_CONTEXT_DIR}/onboard.log. The old bash runner used to redirect onboarding output there; the typed orchestrator captured it under .e2e/actions/<action-id>.log. Fix: add optional aliasPath to PhaseAction; compiler sets aliasPath to 'onboard.log' for the onboarding profile action; orchestrator copies the action evidence log to the alias on success. Best- effort - alias copy failures do not fail the action. Live evidence from run 26550936453 (canonical ubuntu-repo-cloud-openclaw): - environment.install.repo-current: passed in 14.2s - onboarding.profile.cloud-openclaw: passed in 302s (real onboarding!) - onboarding.base.cli-installed: passed - onboarding.preflight.passed: failed (onboard.log not found) <- fixed - runtime.* (10 steps): all 'missing key(s)' <- fixed by #1 Tests: 38/38 phase-orchestrator (was 36; +2 alias tests), 294/294 scenario framework. shellcheck clean. Validation gate next: redispatch e2e-scenarios-all and confirm runtime steps actually exercise the SUT (real pass/fail, not key errors). Signed-off-by: Julie Yaunches <jyaunches@nvidia.com>

Addresses PR Review Advisor finding #1: 'Secret-bearing child process output is persisted and uploaded without redaction'. Designed in the spirit of the test/e2e-scenario architecture: secret hygiene is FRAMEWORK INFRASTRUCTURE, not a per-script / per-action / per-workflow concern. Same one-mode discipline that motivates the rest of this PR (no flag, no env var, no helper bypasses redaction). New module: test/e2e-scenario/scenarios/orchestrators/redaction.ts Sits next to context.ts, owns: - redactString(text): canonical token-shape redaction - buildChildEnv(base, opts): minimal allowlisted env + declared secretEnv passthrough only - pipeRedacted(src, log): single I/O wrapper used by both runAction and runShellStep Pattern set is a framework-local mirror of src/lib/security/secret-patterns.ts. The framework deliberately does NOT import from src/lib/security/ to keep the framework-vs-product boundary clean and avoid cross-tsconfig ESM issues. A new parity test (e2e-redaction-parity.test.ts) asserts the two pattern sets stay in lockstep so adding a token shape in one place keeps both layers honest. Typed contract extension: PhaseAction.secretEnv?: readonly string[] AssertionStep.secretEnv?: readonly string[] Compiler decides which secrets cross the framework boundary, the same way aliasPath / scriptRef / timeoutSeconds are declared metadata. buildChildEnv enforces a 'secret-key shape' contract on every entry (must end with API_KEY/TOKEN/SECRET/PASSWORD/CREDENTIAL/ PASSPHRASE/PRIVATE_KEY/ACCESS_KEY) so the secretEnv channel cannot silently allowlist non-secret variables. compiler.ts: declares onboard secretEnv per profile cloud-* profiles -> ['NVIDIA_API_KEY'] local-ollama-openclaw -> [] Install actions declare nothing (the installer itself does not authenticate to the cloud). phase.ts: two call sites updated, identical pattern in both runAction: buildChildEnv(...) + pipeRedacted(stdout, log) + pipeRedacted(stderr, log, tail<-redactedChunk) runShellStep: same shape Spawn-error 'message' fields are also wrapped with redactString as defense-in-depth. Workflow upload tightening (.github/workflows/e2e-scenarios.yaml): - include-hidden-files: false Hidden dotfiles under the workspace can carry raw secrets (notably .e2e/context.env, written by e2e_context_set without redaction). Diagnostic dumps of context use e2e_context_dump which redacts on emit. - path: explicit subpath list (.e2e/actions/, .e2e/logs/, result JSONs, plan/run-plan, onboard.log) instead of blanket .e2e/. workflow-boundary.mts contract test updated to assert the new invariant; e2e-scenarios-workflow.test.ts expectations updated. Bash side: unchanged. test/e2e-scenario/runtime/lib/context.sh:: e2e_context_dump already redacts via _e2e_context_is_sensitive_key. This module covers the TS-spawned-child path that lacked coverage. Tests (e2e-phase-orchestrators.test.ts, +5): - test_should_not_persist_secret_shaped_child_output_into_evidence Real shell child writes nvapi-, ghp_, sk-, xoxb-, Bearer ... Asserts evidence log, phase result.json, and result.message contain none of the literal tokens but do contain <REDACTED>. - test_should_drop_non_allowlisted_parent_env_unless_declared_in_secretEnv Sets a sentinel parent-env var, runs child without declaring it, asserts child cannot see it. PATH and E2E_PHASE still arrive. - test_should_pass_declared_secretEnv_through_to_child Declares the var in step.secretEnv, asserts child sees it. - test_should_reject_non_secret_shaped_keys_in_secretEnv_at_runtime buildChildEnv throws on FOO_VAR; the secret-key-shape contract keeps the allowlist boundary honest. - test_should_declare_NVIDIA_API_KEY_only_for_cloud_onboarding_actions Compiler-level contract: cloud onboard declares NVIDIA_API_KEY, local-ollama declares nothing. Tests (e2e-redaction-parity.test.ts, new file, +2): - TOKEN_PREFIX_PATTERNS framework copy matches product source - CONTEXT_PATTERNS framework copy matches product source 308/308 e2e-scenario framework tests pass. Signed-off-by: Julie Yaunches <jyaunches@nvidia.com>

ci: add GitHub Actions workflow to auto-update release notes on push …

c620af8

…to main Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Signed-off-by: Aaron Erickson <aerickson@nvidia.com>

cluster2600 mentioned this pull request Mar 18, 2026

Mission Control cannot connect to OpenClaw gateway inside NemoClaw sandbox #328

Closed

wscurran added the CI/CD Use this label to identify issues with NemoClaw CI/CD pipeline or GitHub Actions. label Mar 20, 2026

wscurran added the Integration: GitHub Use this label to identify GitHub integration issues with NemoClaw. label Mar 20, 2026

ericksoa closed this Mar 22, 2026

This was referenced Mar 22, 2026

nemoclaw fails to detect GPU on NVIDIA Jetson Thor and Orin #300

Closed

Install issues #360

Closed

zNeill mentioned this pull request Mar 26, 2026

[Linux] nemoclaw deploy fails: Brev CLI does not support --gpu flag #998

Closed

ACCGAGTT mentioned this pull request Mar 27, 2026

Web search tools fail in sandbox: DNS resolution blocked (getaddrinfo EAI_AGAIN) #626

Closed

drmarcopapa mentioned this pull request Mar 31, 2026

npm install inside sandbox extremely slow — 310s timeouts on FETCH_ERROR retries #1043

Closed

2 tasks

prekshivyas mentioned this pull request Apr 6, 2026

fix(preflight): warn when installed OpenShell version differs from tested version #1314

Closed

Dongni-Yang mentioned this pull request Apr 7, 2026

fix(sandbox): guard openclaw configure and export gateway token (#1114) #1305

Merged

ColinM-sys mentioned this pull request Apr 9, 2026

docs(k8s): document evaluation-only patterns and production alternatives (#1442) #1676

Closed

3 tasks

ericksoa mentioned this pull request Apr 12, 2026

feat(vm): add openshell-vm microVM gateway backend (opt-in via NEMOCLAW_GATEWAY_BACKEND=vm) #1791

Closed

10 tasks

prekshivyas mentioned this pull request Apr 16, 2026

fix(cli): close stale-lock cleanup race in acquireOnboardLock (#1281) #1656

Merged

5 tasks

ericksoa mentioned this pull request Apr 17, 2026

fix(onboard): preserve user's policy preset selections across sandbox… #1997

Merged

13 tasks

This was referenced May 7, 2026

fix(preflight): replace reinvented OpenClaw health check with NemoClaw CLI calls #3208

Closed

Epic: unify warnings, advisories, and fatal exits under a single registry #3213

Open

feat(e2e): introduce scenario-based setup matrix and runner #3290

Closed

prekshivyas mentioned this pull request May 11, 2026

[All Platforms][Regression] node version is not upgrading #2178

Closed

github-actions Bot mentioned this pull request May 13, 2026

feat(e2e): scenario-based setup matrix + Phase 1 migration infrastructure #3363

Merged

12 tasks

raymondlo84 mentioned this pull request May 13, 2026

Jetson Thor: Failed to Install Nemoclaw during OnBoard #3473

Closed

2 tasks

cjagwani mentioned this pull request May 14, 2026

[All Platforms][Onboard] ./install.sh and nemoclaw onboard enter dead loop — sandbox can't reach gateway, "firewall" hint is misleading #3456

Closed

cjagwani mentioned this pull request May 14, 2026

fix(onboard,uninstall): replace misleading recovery messages (#3456) #3520

Merged

cjagwani mentioned this pull request May 14, 2026

[Linux][Uninstall] nemoclaw uninstall leaves ~/.local/state/nemoclaw/ behind (residual of #3456 #5c) #3535

Closed

3 tasks

wangericnv mentioned this pull request May 15, 2026

[Ubuntu 24.04][Sandbox] openshell sandbox download produces a directory instead of file content + does not reject path-traversal targets #3345

Closed

hulynn mentioned this pull request May 18, 2026

[macOS][Onboard] back/exit keywords swallowed as literal API key at provider prompts #3697

Closed

wangericnv mentioned this pull request May 20, 2026

[DGX Station][Inference] NIM-local onboard fails — docker pull nvcr.io/nim/nvidia/nemotron-3-nano:latest returns "error from registry: Incorrect Repository Format" #3885

Open

hulynn mentioned this pull request May 20, 2026

[Nemoclaw] [All Platforms]Hermes onboarding: sandbox build succeeds but sandbox never reaches Ready and is deleted after 180s timeout #3764

Closed

github-actions Bot mentioned this pull request May 20, 2026

fix(discord): route gateway through sandbox proxy #3935

Merged

sandl99 mentioned this pull request May 22, 2026

[Messaging] Add serializable plan applier and reachability hook boundary #3996

Closed

wangericnv mentioned this pull request May 22, 2026

chore: upgrade agent runtime dependencies #3925

Merged

cjagwani mentioned this pull request May 22, 2026

refactor(cli): extract provider inference handlers #3871

Merged

12 tasks

hulynn mentioned this pull request May 28, 2026

[All Platforms][Docs] quickstart.mdx wizard description still out of sync with v0.0.53 (follow-up of NVB 6187477) #4417

Closed

teknium1 mentioned this pull request May 29, 2026

fix(hermes): ship stock Hermes default soul + move env info to env-hint #4524

Merged

PrachiShevate-nv mentioned this pull request May 29, 2026

[Ubuntu 24.04][CLI] nemoclaw hosts-list fails with Docker unknown shorthand flag #4317

Open

Dongni-Yang mentioned this pull request Jun 2, 2026

fix(onboard): don't leave a default sandbox when cancelling at policy presets #4642

Merged

4 tasks

wscurran added area: ci CI workflows, checks, release automation, or GitHub Actions area: integrations Third-party service integration behavior chore Build, CI, dependency, or tooling maintenance labels Jun 3, 2026

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

ci: auto-update release notes on push to main#1

ci: auto-update release notes on push to main#1
ericksoa wants to merge 1 commit into
mainfrom
ci/auto-release-notes

ericksoa commented Mar 16, 2026

Uh oh!

wscurran commented Mar 20, 2026

Uh oh!

cv commented Mar 21, 2026

Uh oh!

ericksoa commented Mar 22, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

ericksoa commented Mar 16, 2026

Summary

How it works

Test plan

Uh oh!

wscurran commented Mar 20, 2026

Uh oh!

cv commented Mar 21, 2026

Uh oh!

ericksoa commented Mar 22, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants