Loading...
Loading...
Found 1,573 Skills
Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).
Score and compare images using vision LLMs as judges. YAML-defined criteria presets for 11 use cases (text-to-image, photorealism, document OCR, charts, UI, portrait, product, scientific, invoice, alt-text, artistic style). Supports OpenAI, Anthropic, Gemini, Mistral, and OpenRouter as judge providers. Keys auto-decrypted via SOPS + age.
Use when writing, reviewing, or committing code to enforce Karpathy's 4 coding principles — surface assumptions before coding, keep it simple, make surgical changes, define verifiable goals. Triggers on "review my diff", "check complexity", "am I overcomplicating this", "karpathy check", "before I commit", or any code quality concern where the LLM might be overcoding.
AI autonomous research agent for LLM training optimization using opencode as the agent. The agent autonomously modifies train.py, runs experiments, evaluates val_bpb, and iterates to find the best model. Use when: "run autoresearch", "start experiment", "train model", "autonomous research", "optimize LLM training".
Return public original model architecture diagrams for user-specified LLM, VLM, MoE, diffusion, OCR, and SGLang/sgl-cookbook model families. Use when the user asks for a model structure chart, architecture diagram, or rendered image link for a specific model such as DeepSeek, GLM, Qwen, Kimi, MiniMax, Step, Hunyuan, or Qwen3-VL.
Use when writing or editing a system prompt for any LLM API or SDK (any code passing a `system=` / `system` role parameter, or a `.txt`/`.md` file holding such a prompt). Applies prompt-engineering and prompt-caching best practices.
Track, optimize, and control token consumption across multi-agent systems. Covers budget allocation, real-time monitoring, cost attribution, per-agent limits, and proactive cost optimization for production LLM deployments.
Route low-risk coding tasks to cheaper LLMs while keeping Codex for high-risk decisions, using MCP tools for cost-aware delegation
Guides AI ops leadership—LLM SRE, model/prompt releases, eval/incidents, cost/capacity, vendors, and cross-functional cadence. Use for AI platform ops, LLM SLAs, incidents, rollout governance, unit economics, red-team/eval gates, and team rituals—not memory (ai-memory-developer), context code (ai-context-engineer), security programs (cybersecurity), token roadmaps (ai-token-improvement-plan-engineer), solution architecture (applied-ai-architect-commercial-enterprise), skills portfolio (ai-skill-manager), or vertical AI product eng management (engineering-manager-vertical-ai-products). Prompt/eval team management and golden-set release policy: engineering-manager-agent-prompts-evals. Safeguard inference platform: ml-infrastructure-engineer-safeguards. Safeguard model research: ml-research-engineer-safeguards.
Design, test, and optimize prompts for LLM interactions. Cover prompt patterns (few-shot, chain-of-thought, ReAct), system prompt design, output formatting, prompt evaluation, and prompt optimization techniques. Triggers on "write prompt", "optimize prompt", "design system prompt", "few-shot examples", "chain of thought", "prompt evaluation", "LLM output formatting", "prompt testing", or "prompt patterns".
Autonomously set up an OpenClaw bot on a fresh Yandex Cloud VM in Kazakhstan (kz1-a, Karaganda). Asks the user for exactly two things — a Telegram bot token and one of three LLM access options (Anthropic API key, OpenRouter API key, or OpenAI Codex OAuth via ChatGPT Plus/Pro subscription) — then handles VM creation, hardening, OpenClaw install, CEO AI OS workspace seeding, Telegram pairing, chat_id auto-detection, and bot-reply verification on its own. The only other actions the user performs are pressing /start in Telegram once and (if Codex) confirming a device code on auth.openai.com. Use when the user says install OpenClaw to Yandex Cloud, deploy OpenClaw to YC Kazakhstan, set up my CEO bot in YC KZ, I am at OpenClaw workshop and need my own bot, create a Yandex Cloud VM for OpenClaw, or any close paraphrase. Targets a ~15-minute end-to-end run for non-DevOps users (founders, CEOs, marketing leads). Supports two modes of accessing Yandex Cloud — Plan A (the user's own YC Kazakhstan account via OAuth) and Plan B (a workshop-key bundle provided by the workshop organizer, for participants without their own YC account). The mode is auto-detected from the inputs. For local-machine OpenClaw install, use openclaw/install.sh in this repo instead. Companion skill openclaw-guide is required; prepare-yc-workshop is the matching organizer-side skill that produces the bundles consumed in Plan B; openclaw-user-onboarding is auto-invoked after Step 5 to collect the five basic facts about the user (identity, focus, style, tools, anti-patterns) and write them into USER.md so the bot is useful from message one.
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.