[codex] disable unsupported OpenClaw self-update hints in sandbox#1215
[codex] disable unsupported OpenClaw self-update hints in sandbox#121513ernkastel wants to merge 4 commits intoNVIDIA:mainfrom
Conversation
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (1)
✅ Files skipped from review due to trivial changes (1)
📝 WalkthroughWalkthroughBuild-time config now sets Changes
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~20 minutes Poem
🚥 Pre-merge checks | ✅ 3✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Comment |
|
✨ Thanks for submitting this pull request, which proposes a way to disable unsupported OpenClaw self-update hints in sandbox. Possibly related open issues:
|
|
Temporarily closing this bundled security PR to free one contributor PR slot for the grouped cleanup PR. I’ll link the replacement PR here as soon as it is open. |
…names (#1416) ## Summary Bundles the remaining sandbox command-hardening work with the Telegram fail-closed cleanup and the unsupported self-update-hint fix. This now includes the original `#1416` scope plus the changes that had temporarily been split into `#1500`. `#1499` remains separate on purpose. ## Linked Issues - Fixes #1029 ## Related PRs / Issues - follow-up to `#1392` - folds in `#1218` - folds in `#1215` - replaces `#1500` - keeps `#1499` separate - addresses `#896` ## Changes - re-validates sandbox names at the `createSandbox()` boundary and removes the remaining shell-string dependency from follow-on sandbox command paths - adds `runFile()` and uses argv-style execution for `setup-dns-proxy.sh` - replaces the dashboard readiness probe with the structured OpenShell helper path - requires an explicit Telegram chat allowlist before the bridge forwards prompts - adds `nemoclaw telegram` subcommands and `nemoclaw start --discover-chat-id` - preserves the reserved-sandbox-name guard added during the Telegram review follow-up - disables unsupported OpenClaw self-update hints in the generated sandbox config - propagates saved Telegram allowlists into the remote deploy env so deployed bridges stay fail-closed too - updates focused CLI/deploy tests to match the current services-based startup path on `main` ## Why These changes all tighten the default security posture around operator-managed sandboxes: - sandbox creation and follow-on helper execution rely less on shell-string construction - Telegram bridge access now fails closed unless the operator explicitly allowlists chat IDs - sandbox images stop advertising an unsupported in-container self-update path Keeping them together in `#1416` makes the remaining security review surface smaller while still leaving the separate immutable-hardening follow-up in `#1499` alone. ## Validation - `npm run build:cli` - `npx vitest run src/lib/deploy.test.ts src/lib/onboard-session.test.ts test/onboard.test.js test/cli.test.js test/runner.test.js test/service-env.test.js test/registry.test.js test/shellquote-sandbox.test.js` ## Risks / Notes - `npm run typecheck:cli` still hits the repo's existing `src/lib/*.test.ts -> ../../dist/lib/*` type-resolution issue in this environment, so validation here relies on the targeted build plus Vitest coverage above - `#1499` remains separate on purpose Signed-off-by: Chia Min Jun Lennon <LennonCMJ@live.com> --------- Signed-off-by: latenighthackathon <latenighthackathon@users.noreply.github.com> Signed-off-by: 13ernkastel <LennonCMJ@live.com> Co-authored-by: latenighthackathon <latenighthackathon@users.noreply.github.com> Co-authored-by: Test User <test@example.com> Co-authored-by: Carlos Villela <cvillela@nvidia.com>
Summary
This PR stops NemoClaw sandbox images from advertising
openclaw updateas the supported upgrade path.Root cause
NemoClaw installs
openclawinto the sandbox image at build time and pins that version inDockerfile.base, but the generated sandbox config still allows OpenClaw's startup update hint. Inside the sandbox that hint tells users to runopenclaw update, which leads them into an in-container self-update flow that is not how NemoClaw images are upgraded.What changed
openclaw.jsonupdate.checkOnStarttofalseUser impact
Users inside a NemoClaw sandbox will stop seeing a misleading recommendation to run
openclaw update. The supported path remains shipping a newer pinned OpenClaw version in the image and recreating the sandbox from that image.Addresses #1029.
Validation
bash -n test/e2e-gateway-isolation.shgit diff --check./test/e2e-gateway-isolation.sh(dockeris not installed in the publishing environment)Summary by CodeRabbit
Documentation
Configuration
Tests