Issues

• Hermes: reading config.yaml puts an openshell:resolve:env placeholder in the next request body; OpenShell proxy resets the connection → every turn fails with 'Connection error'

  NVIDIA/NemoClaw#14056 · Filter by author jcoutsousa opened on Nov 2, 2026

• Planning doesn't catch a model field the image's adapter doesn't accept

  v1
  NVIDIA/NemoClaw#14055 · Filter by author cv opened on Nov 1, 2026

• Proposal: global nemohermes doctor (host readiness without a registered sandbox) + interrupted-onboarding surfacing

  NVIDIA/NemoClaw#14040 · Filter by author Soloh89 opened on Oct 31, 2026

• Require structured telemetry labels for all QA runs

  area: e2earea: observability
  Task

  NVIDIA/NemoClaw#14035 · Filter by author wscurran opened on Oct 30, 2026 · Roadmap: Q4 2026
  1

  

• Check real v1 telemetry records in TEST on Linux and macOS

  area: e2earea: observabilityplatform: linuxplatform: macosv1
  Task

  NVIDIA/NemoClaw#14033 · Filter by author wscurran opened on Oct 30, 2026 · Roadmap: Q4 2026
  

• Send telemetry from apply and destroy, and document it

  area: cliarea: docsarea: observabilityv1
  Task

  NVIDIA/NemoClaw#14032 · Filter by author wscurran opened on Oct 30, 2026 · Roadmap: Q4 2026
  

• Build the v1 telemetry sender in nemoclaw-cli

  area: cliarea: observabilityv1
  Task

  NVIDIA/NemoClaw#14031 · Filter by author wscurran opened on Oct 30, 2026 · Roadmap: Q4 2026
  

• Confirm the TEST receiver accepts a v1-style telemetry record

  area: observabilityv1
  Task

  NVIDIA/NemoClaw#14029 · Filter by author wscurran opened on Oct 30, 2026 · Roadmap: Q4 2026
  

• Stop untagged Classic builds from reporting version 0.1.x

  area: observabilityarea: packagingchore
  Task

  NVIDIA/NemoClaw#14026 · Filter by author wscurran opened on Oct 30, 2026 · Roadmap: Q4 2026
  1

  

• [Ubuntu 22.04][Agent&Skills] Hermes inference still returns HTTP 401 after Discord enrollment rebuild

  area: providersarea: securityintegration: discordintegration: hermesneeds: triageNV QAprovider: nvidia
  Bug

  NVIDIA/NemoClaw#14020 · Filter by author hulynn opened on Oct 30, 2026
  

• [Ubuntu 24.04][Agent&Skills] Hermes status says credentials invalid and inference returns 401 after rebuild

  area: integrationsarea: providersarea: securityintegration: hermesneeds: triageNV QAprovider: nvidia
  Bug

  NVIDIA/NemoClaw#14019 · Filter by author hulynn opened on Oct 30, 2026

• [64GB DGX Spark][Docs] Set Up vLLM: extra serve args switch a 63 GB Spark to the 64 GB+ recipe

  area: onboardingarea: providersneeds: triageNV QAplatform: dgx-sparkprovider: vllm
  Bug

  NVIDIA/NemoClaw#14016 · Filter by author hulynn opened on Oct 30, 2026

• [64GB DGX Spark][Docs] Set Up vLLM: example with --disable-log-requests fails (unrecognized argument)

  area: docsarea: providersneeds: triageNV QAprovider: vllm
  Bug

  NVIDIA/NemoClaw#14015 · Filter by author hulynn opened on Oct 30, 2026

• [Ubuntu 24.04][Onboard] OpenAI-compatible endpoint onboarding fails after Ready with "did not return one exact durable sandbox identity"; no sandbox exists and destroy cannot clear it

  area: providersarea: sandboxneeds: triageNV QAprovider: openaiUAT
  Bug

  NVIDIA/NemoClaw#14013 · Filter by author haoyingw opened on Oct 30, 2026

• [DGX Station][Sandbox] onboard cannot restore provider snapshot acceleration state when converting a CPU-only sandbox to GPU

  area: sandboxneeds: triageNV QA
  Bug

  NVIDIA/NemoClaw#14010 · Filter by author mercl-lau opened on Oct 30, 2026

• [DGX Station][Install] installer reports a stalled github.com lookup as "Could not verify install ref" instead of a timeout

  area: install
  Bug

  NVIDIA/NemoClaw#14007 · Filter by author wangericnv opened on Oct 30, 2026
  

• Report sandbox service URLs in apply output

  area: cliarea: sandboxv1
  Enhancement

  NVIDIA/NemoClaw#13996 · Filter by author kaofelix opened on Oct 29, 2026

• [Epic] NemoClaw v1 operation telemetry

  area: cliarea: observabilityfeaturev1
  Epic

  NVIDIA/NemoClaw#13995 · Filter by author wscurran opened on Oct 29, 2026 · Roadmap: Q4 2026

• [V1 E2E] Cache model files and weights for local-model runs

  area: ciarea: e2earea: local-modelsneeds: triagev1
  Task

  NVIDIA/NemoClaw#13994 · Filter by author rsliter opened on Oct 29, 2026

• [All Platforms][Onboard] Re-onboarding a sandbox that records direct tool disclosure fails unless --tool-disclosure is passed again

  area: onboardingarea: sandbox
  Bug

  NVIDIA/NemoClaw#13975 · Filter by author rluo8 opened on Oct 29, 2026

• [WSL2][Onboard] onboard --resume switches the recorded NVIDIA Endpoints provider to Ollama and keeps the cloud model name

  area: onboardingarea: providersneeds: triageNV QAplatform: wslprovider: nvidiaprovider: ollama
  Bug

  NVIDIA/NemoClaw#13968 · Filter by author mercl-lau opened on Oct 29, 2026

• [Ubuntu 22.04][Sandbox] restarting a NemoHermes sandbox container does not relaunch the agent; health never recovers

  area: sandboxintegration: hermesneeds: triageNV QAplatform: container
  Bug

  NVIDIA/NemoClaw#13967 · Filter by author mercl-lau opened on Oct 29, 2026
  1

• Kubernetes destroy waits the termination grace period for every sandbox

  area: performancearea: sandboxplatform: k8sv1
  Bug

  NVIDIA/NemoClaw#13959 · Filter by author cv opened on Oct 29, 2026

• nemoclaw launch fails after mcp add: provider-composed _provider_* policy entry rejected by launch readiness validation

  area: policyarea: sandboxarea: securityintegration: openclaw
  Bug

  NVIDIA/NemoClaw#13956 · Filter by author scott01272001 opened on Oct 29, 2026
  

• A gateway restart sometimes fails to start a Docker sandbox again

  area: sandboxplatform: containerv1
  Bug

  NVIDIA/NemoClaw#13953 · Filter by author cv opened on Oct 29, 2026
  2