Skills
96,770 packaged capabilities and playbooks for agents. Install any of them with one command.
Agent workflows
Design and audit AI model routers that balance task quality, latency, cost, reliability, safety, privacy, context size, tool use, fallback, caching, evals, budgets, and provider outages. Use when an agent or product must choose models dynamically instead of hard-coding one default model.
Automatically activates when over-engineering patterns are detected: (1) Modifying code or files the user did not explicitly ask to change (2) Creating new abstraction layers (class, interface, factory, wrapper) without being asked (3) Adding comments, documentation, JSDoc, or type annotations without being asked (4) Introducing new dependencies without being asked (5) Rewriting entire files instead of making minimal edits (6) Diff scope significantly exceeding the user's request (7) User signals like "too much", "don't change that", "only change X", "keep it simple", "stop" (8) Adding error handling, validation, or defensive code for scenarios that cannot occur (9) Generating tests, configuration scaffolding, or documentation without being asked
Orchestrate multi-agent workflows from a Kiro spec using codex (code) + Gemini (UI), including dispatch/review/state sync via AGENT_STATE.json + PROJECT_PULSE.md; triggers on user says "Start orchestration from spec at <path>", "Run orchestration for <feature>", or mentions multi-agent execution.
Autonomous NeMo-RL research agent workflow for directed hypothesis testing and open-ended discovery. Guides agents through the full experiment lifecycle: understanding recipes and environments, wiring RL or NeMo-gym runs, launching reproducible baselines and iterations, analyzing results, preserving human oversight, and using git plus TSV logs as the research ledger. Do NOT use for: bug fixes, code review, documentation, refactoring, dependency updates, or single-file changes.
Brev instance operating guidance for NeMo-RL agents working in /home/ubuntu/RL with limited workspace disk, a larger /ephemeral volume, and optional /home/ubuntu/RL/.env secrets. Use when running nemo-rl-auto-research campaigns, experiments, training jobs, model or dataset downloads, shared cache-heavy commands, log-producing runs, checkpoint generation, W&B or Hugging Face authenticated workflows, or any workflow that may create large files on Brev.
Manage durable working-session memory for coding agents. Use when a user asks to preserve or recover agent context across disconnects, VS Code restarts, long-running work, handoffs, or any session where important state should be written periodically under the repo's session directory. Do NOT use for: simple questions, short tasks, one-off commands, linting, or code review.
Describes the agent skills shipped with NemoClaw and how to access them by cloning the repository. Use when users ask about AI agent support, coding assistant integration, or the .agents/skills/ directory. Trigger keywords: nemoclaw agent skills, ai coding assistant, cursor, claude code, copilot
Connects NemoClaw to a local inference server. Use when setting up Ollama, vLLM, TensorRT-LLM, NIM, or any OpenAI-compatible local model server with NemoClaw. Trigger keywords - nemoclaw local inference, ollama nemoclaw, vllm nemoclaw, local model server, openai compatible endpoint, switch nemoclaw inference model, change inference runtime, nemoclaw additional model, nemoclaw sub-agent model, openclaw sub-agent, agents.list, sessions_spawn, vlm-demo, nemoclaw tool calling, ollama tool calls, vllm tool-call-parser, raw json in tui, nemoclaw inference options, nemoclaw onboarding providers, nemoclaw inference routing.
Presents a risk framework for every configurable security control in NemoClaw. Use when evaluating security posture, reviewing sandbox security defaults, or assessing control trade-offs. Trigger keywords - nemoclaw security best practices, sandbox security controls risk framework, nemoclaw credential storage, openshell provider, api key security, openclaw security controls, nemoclaw security boundary, prompt injection, tool access control.
Explains how to run NemoClaw on a remote GPU instance, including the deprecated Brev compatibility path and the preferred installer plus onboard flow. Use when deploying NemoClaw to a remote VM, onboarding a Brev instance, or migrating away from the legacy `nemoclaw deploy` wrapper. Trigger keywords - deploy nemoclaw remote gpu, nemoclaw brev cloud deployment, nemoclaw plugins, openclaw plugins, install openclaw plugin, nemoclaw onboard from dockerfile, nemoclaw brev web ui, nemoclaw getting started, brev quickstart, nvidia nemotron agent, nemoclaw sandbox hardening, container security, docker capabilities, process limits.
Installs NemoClaw, launches a sandbox, and runs the first agent prompt. Use when onboarding, installing, or launching a NemoClaw sandbox for the first time. Trigger keywords - nemoclaw quickstart, install nemoclaw openclaw sandbox, nemohermes quickstart, hermes agent nemoclaw, run hermes openshell sandbox, nemoclaw prerequisites, nemoclaw supported platforms, nemoclaw hardware software, nemoclaw windows wsl2 setup, nemoclaw install windows docker desktop.
Explains operational tasks after the quickstart: listing sandboxes, status and health checks, logs, diagnostics, port forwards, multiple sandboxes, credential reset, rebuilds, network presets, upgrades, and uninstall. Trigger keywords - manage nemoclaw sandboxes, nemoclaw status, nemoclaw list, nemoclaw dashboard port, nemoclaw rebuild, nemoclaw upgrade sandboxes, nemoclaw uninstall, sandbox mutability, sandbox runtime configuration, sandbox rebuild, nemoclaw backup, nemoclaw restore, workspace backup, openshell sandbox download upload, nemoclaw messaging channels, nemoclaw telegram, nemoclaw discord, nemoclaw slack, nemoclaw wechat, nemoclaw whatsapp, openshell channel messaging, nemoclaw workspace files, soul.md, user.md, identity.md, agents.md, sandbox persistence.
Inspects sandbox health, traces agent behavior, and diagnoses problems. Use when monitoring a running sandbox, debugging agent issues, or checking sandbox logs. Trigger keywords - monitor nemoclaw sandbox, debug nemoclaw agent issues.
Explains how OpenClaw, OpenShell, and NemoClaw form the ecosystem, NemoClaw's position in the stack, what NemoClaw adds beyond the community sandbox, and when to prefer NemoClaw versus integrating OpenShell and OpenClaw directly. Use when users ask about the relationship between OpenClaw, OpenShell, and NemoClaw, or when to use NemoClaw versus OpenShell. Trigger keywords - nemoclaw ecosystem, openclaw openshell, nemoclaw vs openshell, sandboxed openclaw, how nemoclaw works, nemoclaw sandbox lifecycle blueprint, nemoclaw overview, openclaw always-on assistants, nvidia openshell, nvidia nemotron, nemoclaw release notes, nemoclaw changelog.
Describes the NemoClaw integration layer and blueprint architecture and how they orchestrate compatible agent sandboxes. Use when looking up architecture, agent integration, plugin structure or blueprint design. Trigger keywords — nemoclaw architecture, nemoclaw agent architecture, nemoclaw plugin blueprint structure, nemoclaw vs openshell, which cli, nemoclaw cli, openshell cli, sandbox commands, nemoclaw cli commands, nemoclaw command reference, nemoclaw network policy, sandbox egress control operator approval, nemoclaw troubleshooting, nemoclaw debug sandbox issues.
Generates BYO custom safety policies for NVIDIA Nemotron content-safety guardrails — Nemotron-Content-Safety-Reasoning-4B (text) and multimodal Nemotron-3-Content-Safety. Produces a Markdown policy, JSON taxonomy, and drop-in inference prompts. Maps rough words or an existing policy to V2 categories, adding custom categories or topic-following rules.
NocoBase 2 only; never use in a NocoBase 3 project. Use when users need to inspect or maintain NocoBase core AI prerequisites through nb api, including LLM providers, saved services, CLI-versus-UI configuration routing, chat models, embedding discovery, secure credentials, and dependency-safe changes.
Keep an OpenClaw agent's non-sensitive context (selected memory, MD files, notes, and custom skills) under version control in a separate Git repository for remote review/tweaks. Use when setting up or operating a Git-based workflow to export workspace context, commit changes (possibly split into multiple commits), and push on a schedule (e.g., nightly) without leaking secrets.
Execute a spec-first one-task-at-a-time loop using OpenSpec OPSX. Use when asked to implement tasks from a list with single-task scope, spec artifacts, and sequential verification before archiving. Don't use for ad-hoc multi-task implementation, non-OpenSpec task trackers, or writing the specs themselves from scratch.
Operate Orca-managed worktrees, folder contexts, terminals, repos, automations, artifacts, skill sharing, worktree comments, and Orca's embedded browser through the `orca` CLI. Use when the user says "$orca-cli", "Orca worktree", "child worktree", "spawn codex/claude in a worktree", "read/wait/send Orca terminal", "handoff" / "handover" / "give this to another agent", "Orca browser", "orca artifacts", or "share skills". Prefer it over raw git worktree, ad hoc PTYs, or Computer Use when Orca state is involved. Use Computer Use only when a visible window needs GUI control that a CLI, filesystem, or API cannot do.
Run one dev-loop session as the driver - preflight, select, one spec end to end. Harness-generic - grooming drafts and takes run in fresh native subagents of whatever harness drives, takes in their own worktrees (an explicit invocation request may route takes through a runner file instead); the gate stays yours.
Diagnose OpenViking memory issues by tracing backward from a memory file to its archive memory_diff.json and, when needed, session messages. Read-only; use for incorrect content, wrong paths or owners, missing memories, and unexplained updates or deletes.
Understanding Claude Code's plugin registry structure, installation scopes, and common issues. Use when troubleshooting plugin installation problems, understanding why plugins show as installed incorrectly, or manually fixing registry entries.
Scaffolds per-repository agent context so coding agents share the same issue tracker rules, triage label vocabulary, domain glossary, ADR layout, and handoff conventions. Triggers on: "set up project context", "configure agent docs", "create CONTEXT.md", "setup agent workflow", "agent issue tracker setup", "triage labels", "domain glossary for agents". Use when a repo needs durable context files before planning, triage, debugging, TDD, architecture review, or multi-agent implementation.
Onboard an AI agent into any project — detect stack (web, mobile, backend), research up-to-date docs, run a universal 8-category audit (pages, state, UI, API, data, auth, jobs, integrations) with auto-detection and auto-skip, interview the developer, and generate structured context files with task-based routing so the agent never wastes time exploring the project again. Re-running project-onboard on an already-onboarded repo refreshes changed context, re-runs the feature audit, preserves interview answers, and updates routing without requiring separate update or audit commands. Use this skill whenever the user says "onboard", "onboard again", "rerun onboard", "update onboard", "refresh context", "set up context", "scan this project", "generate AGENTS.md", "generate CLAUDE.md", "map this codebase", "audit features", "what does this project do", or any request to help an AI agent understand a new or existing codebase. Also trigger when the user starts working in an unfamiliar repo or an existing repo with stale `.context/` files and would benefit from structured context generation.
Implement a production-ready LLM query loop / agent loop for AI applications. Use this skill whenever the user wants to add tool calling, ReAct-style reasoning-action-observation cycles, function calling loops, query engines, agent runtimes, tool_result feedback, max-turn exits, or Claude Code-like Agent Loop behavior to their own product or codebase.
Autonomous AI coding loop that ships features while you sleep. Use when the user wants to run Ralph, start an autonomous coding session, work overnight, or let Claude implement multiple user stories autonomously. Triggers on "ralph", "autonomous mode", "overnight coding", "autonomous loop", or requests to implement multiple stories without supervision.
Memory consolidation and defragmentation for long-term memory maintenance. Use when asked to consolidate memories, defrag memory, run REM sleep, clean up memory files, or process session logs into durable memory. Also use periodically during heartbeats for memory maintenance.
Use this in code repositories whenever an agent needs to run terminal commands, inspect files, search code, view Git status or diffs, run builds, tests, linters, type checks, package managers, logs, Docker, Kubernetes, or database commands. This skill makes the agent prefer RTK wrappers to reduce noisy command output and save context tokens. Do not use for non-coding conversations.
Scaffold Codex hooks into a real project after auditing the repo, verifying the live official docs, schemas, and runtime source, inspecting the effective `hooks` feature flag, and enabling it in project or user config if needed. Use when a user wants Codex hooks, `.codex/hooks.json`, managed Codex hook refreshes, repo-local lifecycle hooks, reusable agent scripts, or help with `SessionStart`, `SubagentStart`, `PreToolUse`, `PermissionRequest`, `PostToolUse`, `PreCompact`, `PostCompact`, `UserPromptSubmit`, `SubagentStop`, or `Stop`. Trigger on: Codex hooks, hooks.json, hooks feature flag, codex_hooks legacy alias, .codex/config.toml, managed hook scaffold. Do NOT use for Claude Code hooks, `.claude/settings.json`, Husky-only setups, or non-Codex agents.
Run the self-improvement agent to review this session and the ~/.learnings/ log files. Use this skill whenever the user explicitly asks to review learnings, promote entries to CLAUDE.md, do an end-of-session review, or analyze GitHub PRs/issues for recurring patterns. Also use when the user says "promote", "review learnings", "what have we learned", or "self-improvement". Do NOT use this skill just for logging — logging happens automatically without the skill (see Passive Logging below). This skill is specifically for the *review and promotion* workflow.
Capture errors, corrections, and feature requests so your OpenClaw agent improves over time instead of repeating the same mistakes. Use when commands fail, users correct the agent, a requested capability is missing, external tools break, or a better recurring workflow is discovered.
Re-encode verbose prose into a dense telegraphic register — punctuation as connectives, label frames, verbless assertions — without losing normativity or precision. Use when compressing system prompts, tool/function descriptions, skill bodies, or agent instructions; reducing token count or context bloat; making documentation token-efficient for LLM input; or rewriting text in compressed notation.
Use when complex problems require systematic step-by-step reasoning with ability to revise thoughts, branch into alternative approaches, or dynamically adjust scope. Ideal for multi-stage analysis, design planning, problem decomposition, or tasks with initially unclear scope.
Manages session state transitions per SESSION_ORCHESTRATION.md. Controls state updates at /work phase boundaries, escalated transitions on error, and initialized restoration on session resume. Internal workflow use only. Do NOT load for: user session management, login state, app state handling.
Installs Scalekit, then picks AgentKit or SaaSKit for the project. Use when the user wants to install Scalekit, add the plugin, or decide which kit to use. It does not configure a dashboard connection (that's `setup-agentkit`) or set app login env (that's `setup-saaskit`).
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.