Workspace Clawexpert

Deep NemoClaw analysis — NVIDIA's enterprise multi-tenant OpenClaw fork. Architecture, isolation model, sandbox policies, enterprise deployment patterns, and applicable patterns for Perlantir.

Expert reference for all OpenClaw agent tools — exec, web, subagents, skills, llm-task, browser, thinking, elevated, ACP agents, agent-send, diffs, reactions, slash-commands, lobster, loop-detection, btw, exec-approvals, capability-cookbook, pdf, skills-config.

Complete ClawHub skill ecosystem catalog — community skills, what we have, what we're missing, recommended installs. Use when researching available community skills, planning installs for any agent workspace, auditing coverage gaps, or managing the skill ecosystem across all 7 agents.

Patterns from NVIDIA NemoClaw — security, sandboxing, enterprise agent patterns.

Expert reference for all OpenClaw core concepts — agents, sessions, context, compaction, memory, multi-agent routing, model providers, failover, streaming, queues, pruning, system prompt, timezones, usage tracking, and features.

Community, marketplace, security landscape, and external knowledge about OpenClaw. Updated from live research.

Expert reference for all OpenClaw model providers — Anthropic, OpenAI, Google, Ollama, OpenRouter, GitHub Copilot, and more. Covers model string format (provider/model-id), auth modes (API key / token / OAuth / setup-token), model aliases, prompt caching, cost tracking, and fast mode. Includes our live setup (Anthropic token mode, sonnet-4-6 / opus-4-6 / haiku-4-5).

Complete map of the OpenClaw config schema from source code. THE single source of truth for valid config keys.

Knowledge from Anthropic Claude SDK — auth, streaming, tool use, message structure.

Advanced OpenClaw features — ACP protocol, multi-agent patterns, advanced compaction/memory, remote gateway, OpenShell, LLM-task, BTW, Lobster. Sourced from live docs.openclaw.ai and local repo docs.acp.md.

Deep knowledge extracted from the OpenClaw source code. Updated by reading the actual repo.

Deep expert reference for OpenClaw gateway internals — configuration, heartbeat (every/activeHours/model/target), sandboxing (Docker/SSH/OpenShell), secrets management (SecretRef), authentication (token/OAuth), health monitoring, logging, doctor diagnostics, network model, and security threat model. Includes our live setup (port 18789, loopback bind, token auth, SIGUSR1 reload, Docker sandbox, version 2026.3.13).

Complete OpenClaw plugin SDK analysis — every bundled plugin cataloged, high-value unused plugins identified, plugin architecture explained, configuration reference, and SDK guide for custom plugins.

Complete OpenClaw changelog analysis — version history, config keys by version, breaking changes, upgrade paths. The authoritative reference for version-safe config changes. Covers 2026.1.5 through 2026.3.14 (Unreleased).