test(e2e): consolidate sandbox resource-limit coverage

[cv]

Summary

This consolidates sandbox resource-limit validation into the broader sandbox-operations live E2E and replaces the destructive 5,000-process fork storm with a lightweight runtime probe. Fast support tests now protect resource-limit warning detection and a content-free framed artifact protocol, while the retired selector continues to route existing dispatches to the consolidated target.

Changes

• Move the real nemoclaw connect resource-limit probe into sandbox-operations, covering login and interactive shell limits, denied raises, completion markers, and security-warning absence.
• Remove the dedicated live target, workflow job, scorecard row, and destructive fork-storm assertion.
• Add a resource-limit diagnostic predicate and focused E2E-support tests for the current warning, older warning forms, and unrelated output.
• Frame and filter the connected-shell stream before fixture capture so artifacts retain one complete set of numeric limits, standalone begin/end markers, and fixed diagnostic bits. Embedded markers, out-of-frame fields, duplicate fields, and incomplete frames fail without retaining rejected text.
• Preserve sandbox-rlimits-connect as a selector alias because existing controllers and manual workflow dispatches can still submit it. A direct removal would reject or skip those raw dispatches; workflow-plan, workflow-boundary, and E2E workflow tests protect the route to sandbox-operations.
• Keep a temporary non-live test-path bridge for the trusted base workflow used during this removal PR. It only verifies the selector alias and documents removal after the trusted workflow stops referencing the old job.
• Update parity, reporting, release-evidence, and target-registry contracts for the consolidated ownership.

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 changes internal E2E ownership and dispatch only; supported resource-limit behavior and retired-selector compatibility are unchanged, and published documentation does not name the removed lane.
• 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: Independent Codex Desktop review verified the content-free framed artifact protocol, exactly-once field assertions, and exit-only failure metadata; focused protocol and resource-limit tests pass.
• Non-success, skipped, or missing CI check accepted by maintainer — check name, approval link, and follow-up issue:

Documentation Writer Review

• Documentation writer subagent reviewed the completed changes
• Result: no-docs-needed
• Evidence: The PR changes internal E2E coverage, selector compatibility, and content-free artifact handling without changing supported product behavior. The temporary bridge documents its trusted workflow consumer and removal condition.
• Agent: Codex Desktop

DGX Station Hardware Evidence

• Tested on DGX Station
• Tested commit:
• Station profile/scenario:
• Result:
• Supporting evidence:

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 validate:pr passed after refreshing origin/main when hooks were skipped or unavailable
• Targeted behavior tests pass for the current change set, or tests are marked not applicable above — command/result or justification: On the current head, the exact trusted-workflow bridge passed 1/1, focused E2E-support tests passed 62/62, mock-parity tests passed 9/9, and mock/live parity passed. TypeScript, project membership, repository checks, and 114 semantic E2E-phase tests across 71 files passed. The preceding consolidation head passed 151 focused Vitest tests.
• Applicable broad gate passed — npm test for broad runtime/test-harness changes; npm run check for repo-wide validation/coverage changes — command/result: npm test was attempted but stopped after unrelated host-dependent failures involving the local OpenShell gateway/TLS fixture, runner-label/platform fixture, SSH-agent subprocess cases, and DGX host detection. No changed test failed; all focused suites were rerun with the active SSH agent and passed.
• 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: Carlos Villela cvillela@nvidia.com

[coderabbitai]

Note

Reviews paused

It 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 reviews.auto_review.auto_pause_after_reviewed_commits setting.

Use the following commands to manage reviews:

• @coderabbitai resume to resume automatic reviews.
• @coderabbitai review to trigger a single review.

Use the checkboxes below for quick actions:

• ▶️ Resume reviews
• 🔍 Trigger review

📝 Walkthrough

Walkthrough

The standalone sandbox-rlimits-connect workflow is retired. Its selector maps to sandbox-operations, which performs resource-limit validation. Workflow validators, reports, gates, scorecards, mock parity, and release evidence reflect the consolidated job.

Changes

Sandbox connect consolidation

Layer / File(s)	Summary
Connect resource-limit validation test/e2e/fixtures/resource-limit-diagnostics.ts, test/e2e/live/sandbox-operations.test.ts, test/e2e/support/resource-limit-diagnostics.test.ts, test/e2e/mock-parity.json	Added diagnostic detection, output filtering, and tests for nemoclaw connect shell limits. The sandbox operations test records validation metadata and artifacts.
Workflow selector consolidation tools/e2e/selector-aliases.mts, tools/e2e/workflow-boundary.mts, tools/e2e/sandbox-operations-workflow-boundary.mts, .github/workflows/e2e.yaml, test/e2e/support/e2e-workflow.test.ts, test/e2e/support/workflow-plan.test.ts, test/e2e/support/sandbox-operations-workflow-boundary.test.ts	Mapped the retired selector to sandbox-operations, generated alias-aware workflow conditions, removed standalone-job validation, and added coverage that rejects restoration of the retired job.
Reporting and dispatch updates test/e2e/support/e2e-live-target-gating.test.ts, test/e2e/support/e2e-scorecard.test.ts, tools/e2e/report-e2e-results.mts, test/release-e2e-evidence.test.ts	Removed the retired job from gating, explicit-only reporting, scorecard fixtures, and release dispatch expectations.

Estimated code review effort: 3 (Moderate) | ~20 minutes

Sequence Diagram(s)

sequenceDiagram
  participant E2ESelector as E2E selector
  participant WorkflowBoundary as workflow-boundary
  participant SandboxOperations as sandbox-operations
  participant ConnectedShell as connected shell
  E2ESelector->>WorkflowBoundary: Select sandbox-rlimits-connect
  WorkflowBoundary->>SandboxOperations: Resolve the selector alias
  SandboxOperations->>ConnectedShell: Run connect resource-limit probes
  ConnectedShell-->>SandboxOperations: Return limits and diagnostic status
  SandboxOperations-->>E2ESelector: Return the consolidated E2E result

Loading

Suggested reviewers: aasthajh, afourniernv, ahunnargikar-nvidia

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly and concisely describes the main change: consolidating sandbox resource-limit coverage into the existing E2E target.

✨ Finishing Touches 💡 1

📝 Generate docstrings 💡

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch codex/consolidate-rlimit-e2e

---

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

[github-code-quality]

Code Coverage Overview

Languages: TypeScript

TypeScript / code-coverage/plugin

The overall coverage in commit 84ed851 in the codex/consolidate-rl... branch remains at 96%, unchanged from commit 264564e in the main branch.

TypeScript / code-coverage/cli

The overall coverage in commit 84ed851 in the codex/consolidate-rl... branch remains at 81%, unchanged from commit 264564e in the main branch.

Show a code coverage summary of the most impacted files.

File	main 264564e	codex/consolidate-rl... 84ed851	+/-
src/lib/inferen...ompatibility.ts	95%	94%	-1%
src/lib/sandbox...rce-identity.ts	88%	88%	0%
src/lib/tunnel/services.ts	76%	76%	0%

---

_{Updated August 02, 2026 05:16 UTC}

[coderabbitai]

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/e2e/live/sandbox-operations.test.ts`:
- Around line 40-43: Remove raw connected-shell output from diagnostics and
artifacts in numericProbe, the connect assertion around expectExitZero, and the
code near line 740. Parse and retain only limit values, exit status, markers,
and bounded redacted, content-free metadata; do not include resultText(connect)
or output in assertion messages, logs, or artifacts.

🪄 Autofix (Beta)

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: 4ba287dc-ea2f-4f76-b849-198f083ad235

📥 Commits

Reviewing files that changed from the base of the PR and between 264564e and 574d229.

📒 Files selected for processing (16)

• .github/workflows/e2e.yaml
• test/e2e/fixtures/resource-limit-diagnostics.ts
• test/e2e/live/sandbox-operations.test.ts
• test/e2e/live/sandbox-rlimits-connect.test.ts
• test/e2e/mock-parity.json
• test/e2e/support/e2e-live-target-gating.test.ts
• test/e2e/support/e2e-scorecard.test.ts
• test/e2e/support/e2e-workflow.test.ts
• test/e2e/support/resource-limit-diagnostics.test.ts
• test/e2e/support/sandbox-operations-workflow-boundary.test.ts
• test/e2e/support/workflow-plan.test.ts
• test/release-e2e-evidence.test.ts
• tools/e2e/report-e2e-results.mts
• tools/e2e/sandbox-operations-workflow-boundary.mts
• tools/e2e/selector-aliases.mts
• tools/e2e/workflow-boundary.mts

💤 Files with no reviewable changes (1)

• test/e2e/live/sandbox-rlimits-connect.test.ts

[github-actions]

PR Review Advisor — No blocking findings reported

Advisor assessment: No blocking advisor findings reported
Next action: No advisor follow-up needed.
Findings: 0 blockers · 0 warnings · 0 suggestions

Model lanes

• GPT-5.6 Terra (primary): Completed · high confidence · 0 blockers · 0 warnings · 0 suggestions
• Nemotron 3 Ultra (second opinion): Failed after a partial review · low confidence · 0 blockers · 0 warnings · 0 suggestions

Second-opinion terminology and E2E selections are advisory. They do not change the primary assessment or E2E / PR Gate.

E2E guidance

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

Recommended E2E: cloud-inference, cloud-onboard, security-posture, sandbox-operations, sandbox-rlimits-connect

Workflow run details

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

[coderabbitai]

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/e2e/fixtures/resource-limit-diagnostics.ts`:
- Around line 6-7: Frame resource-limit diagnostics strictly across
test/e2e/fixtures/resource-limit-diagnostics.ts:6-7 by accepting only complete
standalone begin/end marker lines and emitting fields only while inside that
frame; update numericProbe in test/e2e/live/sandbox-operations.test.ts:41-46 to
require each expected field exactly once and reject missing, duplicate, or
out-of-frame fields; replace the embedded-marker test in
test/e2e/support/resource-limit-diagnostics.test.ts:38-58 with coverage
rejecting prefixed markers, out-of-frame fields, and duplicate fields.

🪄 Autofix (Beta)

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: d4b81d26-cb2e-40f2-b8a9-e03d7269f214

📥 Commits

Reviewing files that changed from the base of the PR and between 574d229 and d880400.

📒 Files selected for processing (3)

• test/e2e/fixtures/resource-limit-diagnostics.ts
• test/e2e/live/sandbox-operations.test.ts
• test/e2e/support/resource-limit-diagnostics.test.ts

[coderabbitai]

🧹 Nitpick comments (1)

test/e2e/support/resource-limit-diagnostics.test.ts (1)

79-115: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick win

Add incomplete-frame coverage.

The filter flags an open frame at input close as a protocol error. Add a test with a begin marker and no end marker. Assert resource_limit_protocol_error=1.

As per path instructions, review tests for behavioral confidence rather than implementation lock-in.

🤖 Prompt for 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.

In `@test/e2e/support/resource-limit-diagnostics.test.ts` around lines 79 - 115,
Add an incomplete-frame test alongside the existing filterResourceLimitOutput
protocol tests, passing input containing RESOURCE_LIMIT_CONNECT_BEGIN_MARKER
without RESOURCE_LIMIT_CONNECT_END_MARKER. Assert the result contains
"resource_limit_protocol_error=1\n", focusing on the externally observable
behavior.

Source: Path instructions

🤖 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.

Nitpick comments:
In `@test/e2e/support/resource-limit-diagnostics.test.ts`:
- Around line 79-115: Add an incomplete-frame test alongside the existing
filterResourceLimitOutput protocol tests, passing input containing
RESOURCE_LIMIT_CONNECT_BEGIN_MARKER without RESOURCE_LIMIT_CONNECT_END_MARKER.
Assert the result contains "resource_limit_protocol_error=1\n", focusing on the
externally observable behavior.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 6fdea694-5920-48d4-92c1-de2fcb68bd9d

📥 Commits

Reviewing files that changed from the base of the PR and between d28f766 and d43f127.

📒 Files selected for processing (3)

• test/e2e/fixtures/resource-limit-diagnostics.ts
• test/e2e/live/sandbox-operations.test.ts
• test/e2e/support/resource-limit-diagnostics.test.ts

🚧 Files skipped from review as they are similar to previous changes (1)

• test/e2e/live/sandbox-operations.test.ts