Browse reusable Agent Skills, each with a clear purpose and practical guidance.
Run LLM inference with llama.cpp on CPU, Apple Silicon, AMD/Intel GPUs, or NVIDIA — plus GGUF model conversion and quantization (2–8 bit with K-quants and imatrix). Covers CLI, Python bindings, OpenAI-compatible server, and Ollama/LM Studio integration. Use for edge deployment, M1/M2/M3/M4 Macs, CUDA-less environments, or flexible local quantization.
Play Pokemon games autonomously via headless emulation. Starts a game server, reads structured game state from RAM, makes strategic decisions, and sends button inputs — all from the terminal.
Keeps explicit shell work legible, bounded, and attached to the current workspace thread.
Track ML experiments with automatic logging, visualize training in real-time, optimize hyperparameters with sweeps, and manage model registry with W&B - collaborative MLOps platform
Lark/Feishu CLI skills: lark-cli operations for docs, markdown, sheets, base, calendar, im, mail, task, okr, drive, wiki, slides, whiteboard, apps, approval, attendance, contact, vc, minutes, note, event. Use when the user needs to operate Lark/Feishu resources via lark-cli, send messages, manage documents, spreadsheets, calendars, tasks, OKRs, deploy web pages, or any Feishu/Lark workspace operations.
Drives the KiCad command-line interface (`kicad-cli`, version 9.0) to export schematics and PCBs, produce fabrication outputs, and run design checks. This skill should be used when the user asks to export gerbers, drill, or pick-and-place files, generate a BOM or netlist, run ERC or DRC (including in CI), export a STEP/3D model or PDF/SVG, upgrade KiCad symbol/footprint libraries, or run a KiCad job set. Invoked via "/hardware:use-kicad-cli".
Use BEFORE answering when the user's request concerns the OneClaw/OpenClaw product itself: (1) configuration / setup; (2) troubleshooting or 'why does it behave like this' questions about product runtime; (3) capability boundary — user asks for an action and you are NOT certain the product natively supports it; (4) multi-turn dead-end — after several turns of attempts and the user's problem is still unresolved, stop guessing and consult the manual. LLM common knowledge is not a product fact — answer from official docs, not generic LLM/OS knowledge. **EXCEPTION (you MAY SKIP)**: when another OneClaw skill clearly matches the request, that native skill takes priority. Trigger keywords: API Key, Kimi, provider, model settings, switch model, config file, 设置, 配置, 模型切换, 怎么设置, 在哪配, 故障, 报错, 限制, 为什么. Few-shot triggers: 'Kimi API Key 在哪配？', '怎么换模型？', 'OneClaw 能剪视频吗？'.
Use when making, reviewing, debugging, or scripting Runware API requests. Follow Runware's raw REST/WebSocket request shape, authentication, UUID, async polling, error, retry, and model-selection rules.
Use Tandem Browser's MCP server (local and remote agents) or HTTP API (local and remote agents) to inspect, browse, and interact with the user's shared browser safely. Prefer targeted tabs and sessions, use snapshot refs before raw DOM or JS, verify action completion explicitly, and leave durable handoffs instead of retrying blindly.