fix(cli): restore fresh onboarding device pairing - #7997
Conversation
Signed-off-by: San Dang <sdang@nvidia.com>
sandl99
added
the
area: security
|
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:
📝 WalkthroughWalkthroughThis PR adds restored-clone pairing support across patching, warm-up, approval, identity loading, CLI token selection, fixtures, and regression tests. ChangesRestored-Clone Pairing Support
Estimated code review effort: 3 (Moderate) | ~30 minutes Sequence Diagram(s)sequenceDiagram
participant AutoPairApproval
participant SandboxWarmup
participant OpenClawGateway
participant PatchScript
participant DevicesCLI
AutoPairApproval->>SandboxWarmup: run restored-clone warm-up
SandboxWarmup->>OpenClawGateway: call sessions.create
PatchScript->>OpenClawGateway: apply restored identity handling
PatchScript->>DevicesCLI: select restored paired token
DevicesCLI-->>AutoPairApproval: return paired token
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 |
Code Coverage OverviewLanguages: TypeScript TypeScript / code-coverage/pluginThe overall coverage in commit 4979e10 in the TypeScript / code-coverage/cliThe overall coverage in commit 4979e10 in the Show a code coverage summary of the most impacted files.
Updated |
PR Review Advisor — No blocking findings reportedAdvisor assessment: No blocking advisor findings reported Model lanes
Nemotron output stays in workflow artifacts and does not change the assessment above. E2E guidanceAdvisory only. E2E / PR Gate selects and runs jobs independently. Recommended E2E: 1 warning · 0 suggestionsWarningsWarnings do not block.
|
|
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. |
Signed-off-by: San Dang <sdang@nvidia.com>
Signed-off-by: San Dang <sdang@nvidia.com>
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
src/lib/actions/sandbox/auto-pair-approval.ts (1)
237-240: 🔒 Security & Privacy | 🟡 Minor | ⚡ Quick winClear inherited pairing markers before ordinary approval.
gateway_approval_envremoves onlyOPENCLAW_GATEWAY_*variables. The ordinary path can inherit bothNEMOCLAW_OPENCLAW_FORCE_DEVICE_PAIRINGandNEMOCLAW_OPENCLAW_RESTORED_CLONE_PAIRING. Clear both markers before invoking the approval command.🤖 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 `@src/lib/actions/sandbox/auto-pair-approval.ts` around lines 237 - 240, Update the ordinary approval path that assigns approve_env via gateway_approval_env to remove both NEMOCLAW_OPENCLAW_FORCE_DEVICE_PAIRING and NEMOCLAW_OPENCLAW_RESTORED_CLONE_PAIRING from the inherited environment before invoking the approval command, matching the cleanup already performed in the else branch.
🧹 Nitpick comments (1)
src/lib/actions/sandbox/auto-pair-approval.ts (1)
224-225: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick winConsider extracting the repeated pairing-flag reset into a helper.
The pattern
approve_env.pop('NEMOCLAW_OPENCLAW_FORCE_DEVICE_PAIRING', None)followed byapprove_env.pop('NEMOCLAW_OPENCLAW_RESTORED_CLONE_PAIRING', None)(or the corresponding set) repeats at Line 224-225, Line 238-239, Line 245-246, Line 260-261, and Line 278-279. If a future change adds a third pairing mode or another env var, a missed site can reintroduce mode-confusion between ordinary and restored-clone pairing, the exact class of regression this PR fixes.Extract a small Python helper (e.g.,
def reset_pairing_flags(env): ...) generated once in the template and call it from each site instead of repeating the two-line pop pattern.Also applies to: 237-246, 260-261, 278-279
🤖 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 `@src/lib/actions/sandbox/auto-pair-approval.ts` around lines 224 - 225, Extract the repeated pairing-flag updates in the generated Python template into a single helper, such as reset_pairing_flags(env), that removes both NEMOCLAW_OPENCLAW_FORCE_DEVICE_PAIRING and NEMOCLAW_OPENCLAW_RESTORED_CLONE_PAIRING. Define the helper once and replace each repeated reset or corresponding set sequence at the affected pairing sites with calls to it, preserving the existing ordinary and restored-clone mode behavior.
🤖 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.
Outside diff comments:
In `@src/lib/actions/sandbox/auto-pair-approval.ts`:
- Around line 237-240: Update the ordinary approval path that assigns
approve_env via gateway_approval_env to remove both
NEMOCLAW_OPENCLAW_FORCE_DEVICE_PAIRING and
NEMOCLAW_OPENCLAW_RESTORED_CLONE_PAIRING from the inherited environment before
invoking the approval command, matching the cleanup already performed in the
else branch.
---
Nitpick comments:
In `@src/lib/actions/sandbox/auto-pair-approval.ts`:
- Around line 224-225: Extract the repeated pairing-flag updates in the
generated Python template into a single helper, such as
reset_pairing_flags(env), that removes both
NEMOCLAW_OPENCLAW_FORCE_DEVICE_PAIRING and
NEMOCLAW_OPENCLAW_RESTORED_CLONE_PAIRING. Define the helper once and replace
each repeated reset or corresponding set sequence at the affected pairing sites
with calls to it, preserving the existing ordinary and restored-clone mode
behavior.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: e719d4f5-32c3-4f4b-ad5e-4a02c0062c7d
📒 Files selected for processing (3)
src/lib/actions/sandbox/auto-pair-approval-script.test.tssrc/lib/actions/sandbox/auto-pair-approval.test.tssrc/lib/actions/sandbox/auto-pair-approval.ts
🚧 Files skipped from review as they are similar to previous changes (2)
- src/lib/actions/sandbox/auto-pair-approval-script.test.ts
- src/lib/actions/sandbox/auto-pair-approval.test.ts
Signed-off-by: San Dang <sdang@nvidia.com>
There was a problem hiding this comment.
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
src/lib/actions/sandbox/auto-pair-approval.test.ts (1)
524-528: 🎯 Functional Correctness | 🟡 Minor | ⚡ Quick winMake the pairing modes mutually exclusive in the child environment.
This test sets
NEMOCLAW_OPENCLAW_FORCE_DEVICE_PAIRING, but it also inheritsNEMOCLAW_OPENCLAW_RESTORED_CLONE_PAIRINGfromprocess.env. Since the restored-clone flag now controls forced-authentication detection, both flags can make this test exercise the wrong mode.Remove the restored-clone flag before spawning the child process.
Proposed fix
+ const { + NEMOCLAW_OPENCLAW_RESTORED_CLONE_PAIRING: _restoredClonePairing, + ...baseEnv + } = process.env; return spawnSync("sh", { encoding: "utf-8", input: approvalScript, env: { - ...process.env, + ...baseEnv, PATH: `${tmpDir}:/usr/bin:/bin`, NEMOCLAW_OPENCLAW_FORCE_DEVICE_PAIRING: "1",🤖 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 `@src/lib/actions/sandbox/auto-pair-approval.test.ts` around lines 524 - 528, Remove NEMOCLAW_OPENCLAW_RESTORED_CLONE_PAIRING from the child process environment in the test setup around the environment object containing NEMOCLAW_OPENCLAW_FORCE_DEVICE_PAIRING. Ensure the spawned child inherits no restored-clone flag so the test exercises forced device pairing exclusively.
🤖 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.
Outside diff comments:
In `@src/lib/actions/sandbox/auto-pair-approval.test.ts`:
- Around line 524-528: Remove NEMOCLAW_OPENCLAW_RESTORED_CLONE_PAIRING from the
child process environment in the test setup around the environment object
containing NEMOCLAW_OPENCLAW_FORCE_DEVICE_PAIRING. Ensure the spawned child
inherits no restored-clone flag so the test exercises forced device pairing
exclusively.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: bb32f3e0-9445-4c20-86cf-734b3be0bd41
📒 Files selected for processing (5)
src/lib/actions/sandbox/auto-pair-approval-script.test.tssrc/lib/actions/sandbox/auto-pair-approval.test.tssrc/lib/actions/sandbox/auto-pair-approval.tssrc/lib/actions/sandbox/restore-gateway-pairing.test.tssrc/lib/actions/sandbox/restore-gateway-pairing.ts
🚧 Files skipped from review as they are similar to previous changes (2)
- src/lib/actions/sandbox/auto-pair-approval-script.test.ts
- src/lib/actions/sandbox/auto-pair-approval.ts
Signed-off-by: San Dang <sdang@nvidia.com>
There was a problem hiding this comment.
Actionable comments posted: 1
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
src/lib/actions/sandbox/restore-gateway-pairing.ts (1)
137-146: 🩺 Stability & Availability | 🟠 Major | ⚡ Quick winThe restore entrypoint lacks a bounded retry wrapper for pairing failures.
The snapshot restore function (
runSnapshotRestore) callsestablishRestoredSandboxGatewayPairingexactly once (line 1178–1187 ofsnapshot.ts). When pairing fails, the restore throws an error and directs the user to run${CLI_NAME} ${targetSandbox} connectto retry. This places the retry burden on the user, not on the restore operation itself.
establishRestoredSandboxGatewayPairingcorrectly removes internal retry (per its comment at line 137–138), but the public restore caller does not implement a bounded automatic retry loop. Theconnectcommand provides manual recovery throughrunConnectAutoPairApprovalPass, which is best-effort only and not a bounded pairing reestablishment.Verify that transient pairing-state rotation during restore does not become a user-facing failure. If bounded automatic retries are required by the PR objective, add retry logic to the restore entrypoint or ensure the requirement is documented and deliberately deferred.
🤖 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 `@src/lib/actions/sandbox/restore-gateway-pairing.ts` around lines 137 - 146, Update the runSnapshotRestore caller around establishRestoredSandboxGatewayPairing to perform a bounded automatic retry for transient pairing failures, while keeping establishRestoredSandboxGatewayPairing’s single-attempt behavior unchanged. Retry only the pairing operation, stop after the defined limit, and preserve the existing user-facing failure path when all attempts are exhausted.Source: Path instructions
🧹 Nitpick comments (1)
src/lib/actions/sandbox/auto-pair-warmup.test.ts (1)
73-82: 🎯 Functional Correctness | 🔵 Trivial | 🏗️ Heavy liftAdd behavioral coverage through the public warm-up entrypoint.
This test checks
RESTORED_CLONE_WARMUP_SCRIPTtext only. It does not prove thatrunRestoredSandboxScopeWarmupRunpasses this script toopenshell sandbox execor that the sandbox executessessions.create. A wiring regression can therefore pass these assertions. Add a test through the public entrypoint with a controlled adapter boundary and assert the observable command payload. Keep the string checks as supplemental contract checks.As per path instructions, tests must prefer observable outcomes and prove that public entrypoints reach the replacement path.
🤖 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 `@src/lib/actions/sandbox/auto-pair-warmup.test.ts` around lines 73 - 82, Add behavioral coverage for the public runRestoredSandboxScopeWarmupRun entrypoint using a controlled adapter boundary, asserting that openshell sandbox exec receives the restored-clone warm-up script and its payload performs sessions.create. Retain the existing RESTORED_CLONE_WARMUP_SCRIPT string assertions as supplemental contract checks, while verifying the public entrypoint reaches the replacement execution path.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.
Inline comments:
In `@src/lib/actions/sandbox/auto-pair-warmup.ts`:
- Around line 142-143: Update the restored-clone warm-up command in the
auto-pair flow to export NEMOCLAW_OPENCLAW_RESTORED_CLONE_PAIRING=1 alongside
NEMOCLAW_OPENCLAW_FORCE_DEVICE_PAIRING=1, ensuring the descriptor-backed clone
identity is used. Add a regression assertion verifying this environment variable
is present for the warm-up subprocess.
---
Outside diff comments:
In `@src/lib/actions/sandbox/restore-gateway-pairing.ts`:
- Around line 137-146: Update the runSnapshotRestore caller around
establishRestoredSandboxGatewayPairing to perform a bounded automatic retry for
transient pairing failures, while keeping
establishRestoredSandboxGatewayPairing’s single-attempt behavior unchanged.
Retry only the pairing operation, stop after the defined limit, and preserve the
existing user-facing failure path when all attempts are exhausted.
---
Nitpick comments:
In `@src/lib/actions/sandbox/auto-pair-warmup.test.ts`:
- Around line 73-82: Add behavioral coverage for the public
runRestoredSandboxScopeWarmupRun entrypoint using a controlled adapter boundary,
asserting that openshell sandbox exec receives the restored-clone warm-up script
and its payload performs sessions.create. Retain the existing
RESTORED_CLONE_WARMUP_SCRIPT string assertions as supplemental contract checks,
while verifying the public entrypoint reaches the replacement execution path.
🪄 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: 67763668-42f3-4ac5-814e-add4fd7614b6
📒 Files selected for processing (5)
src/lib/actions/sandbox/auto-pair-approval.tssrc/lib/actions/sandbox/auto-pair-warmup.test.tssrc/lib/actions/sandbox/auto-pair-warmup.tssrc/lib/actions/sandbox/restore-gateway-pairing.test.tssrc/lib/actions/sandbox/restore-gateway-pairing.ts
💤 Files with no reviewable changes (1)
- src/lib/actions/sandbox/restore-gateway-pairing.test.ts
🚧 Files skipped from review as they are similar to previous changes (1)
- src/lib/actions/sandbox/auto-pair-approval.ts
| NEMOCLAW_OPENCLAW_FORCE_DEVICE_PAIRING=1 \\ | ||
| openclaw gateway call sessions.create --params "$params" --json >/dev/null 2>&1 || true |
There was a problem hiding this comment.
🔒 Security & Privacy | 🟠 Major | ⚡ Quick win
🧩 Analysis chain
🏁 Script executed:
#!/usr/bin/env bash
set -euo pipefail
rg -n -C 6 \
'NEMOCLAW_OPENCLAW_(FORCE_DEVICE_PAIRING|RESTORED_CLONE_PAIRING)|RESTORED_CLONE_WARMUP_SCRIPT|runRestoredSandboxScopeWarmupRun' \
.Repository: NVIDIA/NemoClaw
Length of output: 50371
🏁 Script executed:
#!/usr/bin/env bash
set -euo pipefail
printf '%s\n' '--- auto-pair-warmup outline ---'
ast-grep outline src/lib/actions/sandbox/auto-pair-warmup.ts
printf '%s\n' '--- warmup implementation ---'
sed -n '1,215p' src/lib/actions/sandbox/auto-pair-warmup.ts
printf '%s\n' '--- restore pairing callers and environment setup ---'
rg -n -C 10 \
'warmupScopeUpgrade|runRestoredSandboxScopeWarmupRun|NEMOCLAW_OPENCLAW_RESTORED_CLONE_PAIRING|runSandboxWarmupScript|openshell.*run|sandbox exec' \
src/lib/actions/sandbox test \
| head -n 500
printf '%s\n' '--- warmup tests ---'
rg -n -C 12 \
'RESTORED_CLONE_WARMUP_SCRIPT|runRestoredSandboxScopeWarmupRun|warmupScopeUpgrade|auto-pair-warmup' \
src testRepository: NVIDIA/NemoClaw
Length of output: 50372
🏁 Script executed:
#!/usr/bin/env bash
set -euo pipefail
printf '%s\n' '--- warmup tests ---'
sed -n '1,260p' src/lib/actions/sandbox/auto-pair-warmup.test.ts
printf '%s\n' '--- restore pairing action ---'
sed -n '1,240p' src/lib/actions/sandbox/restore-gateway-pairing.ts
printf '%s\n' '--- exact warmup call sites ---'
rg -n -C 18 \
'warmupScopeUpgrade|runRestoredSandboxScopeWarmupRun|runSandboxScopeWarmupRun' \
src/lib --glob '*.ts' --glob '!**/*.test.ts'
printf '%s\n' '--- restored-clone environment construction ---'
rg -n -C 14 \
'NEMOCLAW_OPENCLAW_RESTORED_CLONE_PAIRING|NEMOCLAW_OPENCLAW_IDENTITY_FD|buildAutoPairApprovalScript|approveRestoredClonePairing' \
src/lib/actions/sandbox/restore-gateway-pairing.ts \
src/lib/actions/sandbox/auto-pair-approval.ts \
src/lib/actions/sandbox/*.ts \
| head -n 500Repository: NVIDIA/NemoClaw
Length of output: 50371
🏁 Script executed:
#!/usr/bin/env bash
set -euo pipefail
printf '%s\n' '--- all production assignments and exports ---'
rg -n \
'(^|[[:space:];])export[[:space:]]+NEMOCLAW_OPENCLAW_(FORCE_DEVICE_PAIRING|RESTORED_CLONE_PAIRING)|NEMOCLAW_OPENCLAW_(FORCE_DEVICE_PAIRING|RESTORED_CLONE_PAIRING)[[:space:]]*=' \
src scripts \
--glob '*.ts' --glob '*.mts' --glob '*.sh' --glob '*.py'
printf '%s\n' '--- sandbox environment and startup wiring ---'
rg -n -C 10 \
'proxy-env|PROXY_ENV|sandbox.*env|environment|env=|OPENCLAW_STATE_DIR|NEMOCLAW_OPENCLAW_' \
src/lib/adapters src/lib/actions src/lib/onboard scripts \
--glob '*.ts' --glob '*.mts' --glob '*.py' --glob '*.sh' \
| head -n 700
printf '%s\n' '--- patched identity resolution semantics ---'
sed -n '89,185p' scripts/patch-openclaw-device-self-approval.mts
sed -n '490,525p' scripts/patch-openclaw-device-self-approval.mtsRepository: NVIDIA/NemoClaw
Length of output: 50371
🏁 Script executed:
#!/usr/bin/env bash
set -euo pipefail
printf '%s\n' '--- identity resolver patch ---'
sed -n '115,155p' scripts/patch-openclaw-device-self-approval.mts
sed -n '160,205p' scripts/patch-openclaw-device-self-approval.mts
printf '%s\n' '--- paired-token context patch ---'
sed -n '495,535p' scripts/patch-openclaw-device-self-approval.mts
printf '%s\n' '--- production restored-clone environment evidence ---'
rg -n \
'NEMOCLAW_OPENCLAW_RESTORED_CLONE_PAIRING|NEMOCLAW_OPENCLAW_IDENTITY_FD|NEMOCLAW_OPENCLAW_EXPECTED_DEVICE_ID' \
src scripts \
--glob '*.ts' --glob '*.mts' --glob '*.sh' --glob '*.py' \
| grep -v '/test' \
| head -n 250Repository: NVIDIA/NemoClaw
Length of output: 11528
Set NEMOCLAW_OPENCLAW_RESTORED_CLONE_PAIRING=1 for the restored-clone warm-up. The host process does not export this variable, and the later approval subprocess cannot affect the earlier warm-up. With only NEMOCLAW_OPENCLAW_FORCE_DEVICE_PAIRING=1, OpenClaw resolves the default identity instead of the descriptor-backed clone identity. Add a regression assertion for the environment.
🤖 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 `@src/lib/actions/sandbox/auto-pair-warmup.ts` around lines 142 - 143, Update
the restored-clone warm-up command in the auto-pair flow to export
NEMOCLAW_OPENCLAW_RESTORED_CLONE_PAIRING=1 alongside
NEMOCLAW_OPENCLAW_FORCE_DEVICE_PAIRING=1, ensuring the descriptor-backed clone
identity is used. Add a regression assertion verifying this environment variable
is present for the warm-up subprocess.
Signed-off-by: San Dang <sdang@nvidia.com>
<!-- markdownlint-disable MD041 --> ## Summary Fresh onboarding and restored snapshots now use separate OpenClaw pairing modes. Fresh onboarding keeps the default generated identity path, while restored clones use descriptor-pinned identity and pairing state plus one narrowly bounded recovery when the first verifier publishes the missing scope-upgrade request. This replacement carries forward and credits all six signed commits from #7997, then adds the bounded recovery needed to keep the fresh-onboard and restored-clone E2E paths from regressing one another. ## Related Issue Supersedes #7997 and follows up the pairing regression introduced while fixing #7834. The regression is exercised by `issue-4462-scope-upgrade-approval` and `snapshot-e2e`. ## Changes - Separate ordinary forced pairing from restored-clone descriptor-backed pairing. - Use clone device authentication for the restored-clone write-scope warm-up. - Distinguish an absent clone pending file from malformed or unreadable pairing state. - Retry approval exactly once only after `list-pending-unavailable` and an authenticated `scope-upgrade-pending` verification result. - Keep malformed, ambiguous, mismatched, or unreadable state fail closed. - Add focused unit, patch-upgrade, and live E2E contract coverage for both pairing modes. ## 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 repairs internal OpenClaw pairing orchestration without changing a supported command, configuration, default, or documented workflow. - [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: Codex Desktop reviewed the exact head against all nine security categories and found no blocking or non-blocking findings: #8015 (comment) - [ ] 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: `no-docs-needed` - Evidence: The completed diff changes internal OpenClaw pairing repair behavior only; no public command, configuration, default, error contract, or documentation workflow changed. - Agent: Codex Desktop <!-- docs-review-head-sha: 5be51ee --> <!-- docs-review-agents-blob-sha: c052d60 --> ## DGX Station Hardware Evidence - [ ] Tested on DGX Station - Tested commit: - 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 — 35/35 exact-head focused tests passed; the affected-test run passed 224/224 before the unchanged rebase; `npm run validate:pr` passed on the current head. - [ ] 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` built successfully and passed 23,255 tests locally, with 113 host-dependent failures on macOS/Node 25; required Linux CI and trusted live E2E are authoritative and pending. - [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) --- Signed-off-by: Prekshi Vyas <prekshiv@nvidia.com> <!-- This is an auto-generated comment: release notes by coderabbit.ai --> ## Summary by CodeRabbit * **New Features** * Improved restored-device pairing with dedicated handling for cloned environments. * Added a warm-up step to request required permissions before pairing. * Added clearer status reporting when pending pairing information is unavailable. * **Bug Fixes** * Pairing can now retry when an additional permission approval is needed. * Preserved authentication appropriately across restored-device pairing flows. * Improved compatibility when upgrading existing pairing configurations. * **Tests** * Added coverage for restored-device pairing, retries, unavailable states, and legacy configuration upgrades. <!-- end of auto-generated comment: release notes by coderabbit.ai --> --------- Signed-off-by: San Dang <sdang@nvidia.com> Signed-off-by: Prekshi Vyas <prekshiv@nvidia.com> Co-authored-by: San Dang <sdang@nvidia.com>
3 participants
Summary
The restored-clone pairing path from #7834 made ordinary force pairing require clone descriptors, so fresh onboarding could finish without pairing its CLI device. The restored-clone warm-up could also return before OpenClaw published its pending scope request. This PR separates the two identity modes and publishes one write-scoped request before #7834's one-shot approval. The restored-clone provoke command removes runtime shared auth so OpenClaw uses the clone's stored device credential to publish that request.
Related Issue
Regression follow-up to #7834.
Changes
NEMOCLAW_OPENCLAW_RESTORED_CLONE_PAIRINGmode for the restored-clone approval child. Ordinary force pairing continues to use the default device identity.issue-4462-scope-upgrade-approvalfor deterministic E2E selection.operator.writepairing request withgateway call sessions.createbefore approval. Remove shared gateway credentials only from that provoke child so OpenClaw uses the clone's stored device credential. The coordinator then uses fix(cli): stabilize restored snapshot pairing verification #7834's one strict approval and authenticated verification; it does not add an approval retry or broader scope.Requested Live E2E
snapshot-commandsrestored-pairing-unverifiedafter the strict approval could not list a pending request.restored, verifies gateway pairing, and completes the no-leak lifecycle.Type of Change
Quality Gates
4979e10b1and reported PASS with no findings. Ordinary pairing cannot enter the descriptor-backed clone path. Restored-clone pairing still requires the expected device identity, pinned loopback URL, descriptor-backed state, exact pairing transition, and scoped token. The warm-up removes shared credentials from the provoke child, can publish a request, but cannot approve it. The strict approval and authenticated verifier remain the authorization and success conditions.Documentation Writer Review
no-docs-needed4979e10b1c4de760f1c39bad69a260f77aa181a0againstAGENTS.mdblobc052d60aa2b486bac1235643a56c2d2b1e5c13d5. No findings. Removing shared gateway credentials from this internal provoke child does not change the CLI, configuration, defaults, output, recovery procedure, or supported behavior. Existing restore documentation remains accurate. The reviewer examined 35 focused restore/approval tests, 10 changed tests, CLI typecheck, Biome, and normal hooks.DGX Station Hardware Evidence
scripts/prepare-dgx-station-host.shis unchanged.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 unavailablenpx vitest run --project clifor the six restored-pairing warm-up/approval suites passed 35 tests;npm run test:changedpassed 10 tests; the focused OpenClaw patch integration run on the preceding PR head passed 71 tests with one opt-in real-distribution test skipped;npm run typecheck:cli, Biome, and normal pre-commit/commit-msg/pre-push hooks passed.npm testfor broad runtime/test-harness changes;npm run checkfor repo-wide validation/coverage changes — command/result: An earliernpm testrun reached the 20-minute local bound after failures in untouched host/runtime-dependent gateway, Docker, Hermes image, package-link, and E2E workflow tests. No maintainer waiver is claimed. The changed suites and normal hooks pass, and CI remains authoritative.npm run docsbuilds without warnings (doc changes only)Signed-off-by: San Dang sdang@nvidia.com
Summary by CodeRabbit
New Features
Bug Fixes
Tests