Skip to content

Add nemoclaw-skill (Claude Code skill for NemoClaw) - #4

Open
Koneisto wants to merge 1 commit into
VoltAgent:mainfrom
Koneisto:add-nemoclaw-skill
Open

Add nemoclaw-skill (Claude Code skill for NemoClaw)#4
Koneisto wants to merge 1 commit into
VoltAgent:mainfrom
Koneisto:add-nemoclaw-skill

Conversation

@Koneisto

@Koneisto Koneisto commented Apr 7, 2026

Copy link
Copy Markdown

Adds a new Skills & IDE Integrations section with nemoclaw-skill — a Claude Code skill that encodes NemoClaw operational knowledge.

What it covers

  • 19 critical rules for the OpenShell security boundary (presets, overlay filesystem, SSH bugs, Landlock, native skill design)
  • 6 native skill design principles — one exec call, SOUL.md priority, session snapshot timing
  • Automated workspace recovery — watchdog pattern with sentinel-based detection, workaround for NemoClaw#486
  • Signal bridge production hardening — process detection fixes, watchdog auto-start
  • Trust the architecture — why custom proxies and MCP bridges break, and how to use direct policy entries instead
  • DGX Spark deployment — local vLLM, custom patches, OOM prevention

Also usable as plain Markdown context for any LLM tool, not just Claude Code.

Why a new section?

No existing section fits AI assistant skills/integrations. "Skills & IDE Integrations" is placed between Templates and Example Projects, keeping the document flow logical.

Claude Code skill with 12 critical rules for NemoClaw host management.
Includes automated workspace recovery pattern (workaround for #486),
Signal bridge production hardening, session management, and
multi-sandbox guidance. Also usable as plain Markdown context for
any LLM tool.

https://github.com/Koneisto/nemoclaw-skill
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant