-
Notifications
You must be signed in to change notification settings - Fork 2.9k
Issues
is:issue state:open
is:issue state:open
Issue creation is restricted in this repository
Search results
Install and uninstall a managed OpenShell gateway service for reboot persistence
area: architectureArchitecture, design debt, major refactors, or maintainabilityArchitecture, design debt, major refactors, or maintainabilityarea: installInstall, setup, prerequisites, or uninstall flowInstall, setup, prerequisites, or uninstall flowarea: onboardingOnboarding FSM, provider setup, sandbox launch, or first-run flowOnboarding FSM, provider setup, sandbox launch, or first-run flowarea: packagingPackages, images, registries, installers, or distributionPackages, images, registries, installers, or distributionenhancementNew capability or improvement requestNew capability or improvement requestneeds: designRequires product or architecture directionRequires product or architecture directionneeds: triageAwaiting maintainer classificationAwaiting maintainer classificationplatform: linuxAffects non-Ubuntu Linux environmentsAffects non-Ubuntu Linux environmentsStatus: Open.Add opt-in live MCP tool discovery to mcp status
area: cliCommand line interface, flags, terminal UX, or outputCommand line interface, flags, terminal UX, or outputarea: integrationsThird-party service integration behaviorThird-party service integration behaviorv0.0.84Release targetRelease targetStatus: Open.#6901 In NVIDIA/NemoClaw;[DGX Station][Hermes][Inference]
nemoclaw inference setleaves dashboard chat on previous modelarea: inferenceInference routing, serving, model selection, or outputsInference routing, serving, model selection, or outputsintegration: hermesHermes integration behaviorHermes integration behaviorneeds: triageAwaiting maintainer classificationAwaiting maintainer classificationplatform: arm64Affects ARM64 or aarch64 architectureAffects ARM64 or aarch64 architectureplatform: dgx-stationAffects DGX Station hardware or workflowsAffects DGX Station hardware or workflowsplatform: ubuntuAffects Ubuntu Linux environmentsAffects Ubuntu Linux environmentsStatus: Open.#6893 In NVIDIA/NemoClaw;[All Platforms][CLI&UX] channels status shows Telegram as all-[ok] while the bridge is unreachable — no runtime health probe for Telegram
area: cliCommand line interface, flags, terminal UX, or outputCommand line interface, flags, terminal UX, or outputarea: messagingMessaging channels, bridges, manifests, or channel lifecycleMessaging channels, bridges, manifests, or channel lifecycleintegration: telegramTelegram integration or channel behaviorTelegram integration or channel behaviorplatform: dgx-sparkAffects DGX Spark hardware or workflowsAffects DGX Spark hardware or workflowsplatform: dgx-stationAffects DGX Station hardware or workflowsAffects DGX Station hardware or workflowsv0.0.84Release targetRelease targetVRDCIssues and PRs submitted by NVIDIA VRDC test team.Issues and PRs submitted by NVIDIA VRDC test team.Status: Open.feat: add a core agent identity blueprint contract with Okta and Entra references
area: architectureArchitecture, design debt, major refactors, or maintainabilityArchitecture, design debt, major refactors, or maintainabilityarea: integrationsThird-party service integration behaviorThird-party service integration behaviorarea: policyNetwork policy, egress rules, presets, or sandbox policyNetwork policy, egress rules, presets, or sandbox policyarea: providersInference provider integrations and provider behaviorInference provider integrations and provider behaviorarea: sandboxOpenShell sandbox lifecycle, runtime, config, or recoveryOpenShell sandbox lifecycle, runtime, config, or recoveryarea: securitySecurity controls, permissions, secrets, or hardeningSecurity controls, permissions, secrets, or hardeningenhancementNew capability or improvement requestNew capability or improvement requestneeds: designRequires product or architecture directionRequires product or architecture directionStatus: Open.#6871 In NVIDIA/NemoClaw;[All Platforms][CLI&UX] config set --restart hangs on the gateway restart — config written but never applied to the running agent
area: cliCommand line interface, flags, terminal UX, or outputCommand line interface, flags, terminal UX, or outputintegration: openclawOpenClaw integration behaviorOpenClaw integration behaviorNV QABugs found by the NVIDIA QA TeamBugs found by the NVIDIA QA TeamStatus: Open.[Onboard][Inference] SSRF preflight hard-blocks a legitimately-configured private/internal inference endpoint, with no opt-in override
area: inferenceInference routing, serving, model selection, or outputsInference routing, serving, model selection, or outputsarea: networkingDNS, proxy, TLS, ports, host aliases, or connectivityDNS, proxy, TLS, ports, host aliases, or connectivityarea: onboardingOnboarding FSM, provider setup, sandbox launch, or first-run flowOnboarding FSM, provider setup, sandbox launch, or first-run flowarea: securitySecurity controls, permissions, secrets, or hardeningSecurity controls, permissions, secrets, or hardeningStatus: Open.#6861 In NVIDIA/NemoClaw;[Ubuntu][Inference] managed gateway health probe sends max_tokens to GPT-5 endpoints — probe fails HTTP 400 and gates agent traffic (incomplete fix of #6642)
area: inferenceInference routing, serving, model selection, or outputsInference routing, serving, model selection, or outputsarea: routingRequest routing, policy routing, model selection, or fallback logicRequest routing, policy routing, model selection, or fallback logicintegration: dcodeLangChain Deep Code integration behaviorLangChain Deep Code integration behaviorplatform: ubuntuAffects Ubuntu Linux environmentsAffects Ubuntu Linux environmentsprovider: openaiOpenAI API or OpenAI-compatible provider behaviorOpenAI API or OpenAI-compatible provider behaviorv0.0.84Release targetRelease targetStatus: Open.#6850 In NVIDIA/NemoClaw;[Ubuntu][Onboard] onboard --resume finalizes "OpenClaw is ready" (exit 0) leaving a sandbox with a dead inference route when a committed endpoint is unreachable
area: inferenceInference routing, serving, model selection, or outputsInference routing, serving, model selection, or outputsarea: onboardingOnboarding FSM, provider setup, sandbox launch, or first-run flowOnboarding FSM, provider setup, sandbox launch, or first-run flowv0.0.84Release targetRelease targetStatus: Open.#6849 In NVIDIA/NemoClaw;[Ubuntu][Onboard] loopback (127.0.0.1) custom inference endpoint onboards successfully but the sandbox inference route is HTTP 503 with no unreachable-from-sandbox warning
area: inferenceInference routing, serving, model selection, or outputsInference routing, serving, model selection, or outputsarea: networkingDNS, proxy, TLS, ports, host aliases, or connectivityDNS, proxy, TLS, ports, host aliases, or connectivityarea: onboardingOnboarding FSM, provider setup, sandbox launch, or first-run flowOnboarding FSM, provider setup, sandbox launch, or first-run flowarea: sandboxOpenShell sandbox lifecycle, runtime, config, or recoveryOpenShell sandbox lifecycle, runtime, config, or recoveryplatform: windowsAffects native Windows environmentsAffects native Windows environmentsplatform: wslAffects Windows Subsystem for LinuxAffects Windows Subsystem for Linuxv0.0.84Release targetRelease targetStatus: Open.#6848 In NVIDIA/NemoClaw;nemoclaw statusreports inference healthy on endpoint reachability, not model invocability — a green status can mask an unreachable modelarea: cliCommand line interface, flags, terminal UX, or outputCommand line interface, flags, terminal UX, or outputarea: inferenceInference routing, serving, model selection, or outputsInference routing, serving, model selection, or outputsarea: local-modelsLocal model providers, downloads, launch, or connectivityLocal model providers, downloads, launch, or connectivityarea: providersInference provider integrations and provider behaviorInference provider integrations and provider behaviorStatus: Open.Nemoclaw with hermes teams gateway fails to make tool calls and drops with empty response
area: messagingMessaging channels, bridges, manifests, or channel lifecycleMessaging channels, bridges, manifests, or channel lifecycleintegration: hermesHermes integration behaviorHermes integration behaviorStatus: Open.#6836 In NVIDIA/NemoClaw;