fix(sandbox): normalize staged script permissions

[senthilr-nv]

Summary

NemoClaw now normalizes staged script permissions before Docker consumes either sandbox build context. Fresh installs under a restrictive umask no longer preserve root-only script modes that fail in later non-root image-build stages.

Related Issue

Fixes #7071

Changes

• Normalize the copied scripts/ tree in the legacy sandbox build-context path.
• Normalize stagedScriptsDir after the optimized path copies its required scripts.
• Cover both paths under umask 0077, including a 0700 executable helper and 0700 parent directories.

Type of Change

• 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

• Tests added or updated for changed behavior
• Existing tests cover changed behavior — justification:
• Tests not applicable — justification:
• Docs updated for user-facing behavior changes
• Docs not applicable — justification: this only corrects internal Docker build-context modes; commands, configuration, defaults, APIs, schemas, and documented workflows are unchanged.
• Sensitive paths changed (security, policy, credentials, preflight, onboarding, inference, runner, sandbox, or messaging)
• Sensitive-path review completed or maintainer-approved waiver recorded — reviewer/approval link/justification: requested on this draft before it is marked ready.
• Non-success, skipped, or missing CI check accepted by maintainer — check name, approval link, and follow-up issue:

Verification

• PR description includes a Signed-off-by: line and every commit appears as Verified in GitHub
• Normal pre-commit, commit-msg, and pre-push hooks passed, or npm run check:diff passed when hooks were skipped or unavailable
• Targeted behavior tests pass for the current change set, or tests are marked not applicable above — npx vitest run --project integration test/sandbox-build-context.test.ts: 10 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 run; this is a focused build-context change with direct regression coverage.
• Quality Gates section completed with required justifications or waivers
• No secrets, API keys, or credentials committed
• npm run docs builds without warnings (doc changes only)
• Doc pages follow the style guide (doc changes only)
• New doc pages include SPDX header and frontmatter (new pages only)

---

Signed-off-by: Senthil Ravichandran senthilr@nvidia.com

Summary by CodeRabbit

• Bug Fixes

  • Fixed sandbox builds so staged script files keep readable permissions even when created under restrictive system settings.
  • Improved permissions consistency for staged scripts across both optimized and legacy build staging flows.

• Tests

  • Added/extended test coverage to validate script permissions for both staging modes when using a restrictive umask.
  • Ensures staged scripts, scripts/lib, and scripts/lib/reviewed-npm-archive.mts resolve to the expected readable permissions.

[copy-pr-bot]

Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually.

Contributors can view more details about this message here.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 4bc51303-8b78-4a61-aabc-4ba796a53669

📥 Commits

Reviewing files that changed from the base of the PR and between e2952bf and 2ae2d15.

📒 Files selected for processing (2)

• src/lib/sandbox/build-context.ts
• test/sandbox-build-context.test.ts

🚧 Files skipped from review as they are similar to previous changes (2)

• src/lib/sandbox/build-context.ts
• test/sandbox-build-context.test.ts

---

📝 Walkthrough

Walkthrough

Sandbox build-context staging now normalizes read permissions for scripts in both legacy and optimized paths. Tests create restrictive source modes and verify staged script directories and files remain mode 755 under umask 0077.

Changes

Sandbox script permissions

Layer / File(s)	Summary
Normalize staged scripts src/lib/sandbox/build-context.ts	Both legacy and optimized staging paths apply normalizeReadModesForDockerCopy to staged script content.
Verify restrictive-umask staging test/sandbox-build-context.test.ts	Fixtures use mode 700, and tests verify mode 755 for staged scripts under umask 0077 in both staging paths.

Estimated code review effort: 2 (Simple) | ~10 minutes

Possibly related PRs

• NVIDIA/NemoClaw#6060: Updates sandbox script staging in the same build-context area.
• NVIDIA/NemoClaw#6942: Changes which script files are staged in the sandbox build context.
• NVIDIA/NemoClaw#6967: Changes optimized sandbox staging of script files.

Suggested labels: bug-fix

Suggested reviewers: cv

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly and concisely summarizes the main change: normalizing staged script permissions in sandbox build contexts.
Linked Issues check	✅ Passed	The patch normalizes staged scripts in both legacy and optimized paths and adds umask 0077 regression tests, matching #7071.
Out of Scope Changes check	✅ Passed	The changes are limited to the sandbox build-context fix and related tests; no unrelated scope appears introduced.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

PR Review Advisor — Informational

Advisor assessment: Informational / high confidence
Next action: No advisor follow-up needed.
Findings: 0 blockers · 0 warnings · 0 suggestions
Status: No actionable findings remain in the canonical review ledger.

Model lanes

• GPT-5.6 Terra (primary): Completed · high confidence · 0 blockers · 0 warnings · 0 suggestions
• Nemotron 3 Ultra (second opinion): Completed · high confidence · 0 blockers · 0 warnings · 0 suggestions
• Model comparison: normalized findings match; normalized E2E selections differ; severity counts match.

Nemotron output stays in workflow artifacts and does not change the assessment above.

E2E guidance

Advisory only. E2E / PR Gate selects and runs jobs independently.

Recommended E2E: None

1 optional E2E recommendation

• rebuild-openclaw

Workflow run details

This automated review informs maintainers. Warnings and suggestions do not require a response. A maintainer decides whether to merge.

[senthilr-nv]

Additional Brev validation :

• Tested commit e2952bf.
• Created a fresh checkout under umask 0077; scripts/, scripts/lib/, and scripts/lib/reviewed-npm-archive.mts were all verified as mode 0700 at the source.
• Ran the interactive default OpenClaw Express onboarding through actual BuildKit sandbox creation.
• Build completed all 70 runtime stages. The helper COPY and subsequent stages succeeded, the image exported, the sandbox reached Ready, and the dashboard became live.
• A disposable provider fixture was used only to reach sandbox creation; this was not counted as an inference-quality E2E.
• Cleanup completed: sandbox, provider, mock process, and generated image were removed.

This reproduces the restrictive source-permission precondition and validates the staged-script normalization in a real sandbox build.

[apurvvkumaria]

Security-sensitive maintainer review complete at exact head 2ae2d15. The restrictive-umask fix covers both staging paths; mode normalization preserves executability, adds no write bits, does not follow symlinks, and remains confined beneath the private build context. Exact-head CI, CodeQL, DCO, advisor, and protected E2E are green; no unresolved threads or actionable findings remain.