Total 55,914 skills, AI & Machine Learning has 9310 skills
Showing 12 of 9310 skills
Design task-local harnesses, eval gates, and reusable skill extraction for Claude dynamic workflow mode and other adaptive agent harnesses.
Run team-based orchestration for agent squads using work items, ownership, agent Kanban, merge gates, and control pane handoffs.
Local mirror of OpenAI Codex product documentation (developers.openai.com/codex): CLI, Cloud, web app, IDE extension, hooks, skills, plugins, MCP, subagents, AGENTS.md, prompts, rules, sandboxing, models, pricing, security, and configuration. Use whenever the user asks how Codex behaves, how to install or configure Codex, or what a Codex flag, slash command, or feature does (including informal phrasing such as "hooks", "--resume", "sandbox modes", "cloud environments"). Read this skill's references/ before generic web search for Codex product questions. Do NOT use for Claude Code, Cursor, or other agents -- in particular, do not use for "Claude Code hooks" or general OpenAI API, ChatGPT, Realtime, or non-Codex coding help.
Turn ordinary text plans into rich interactive visual plans with diagrams, file maps, annotated code, open questions, and UI/prototype review when useful.
AI-powered image generation for Salesforce visuals via Nano Banana Pro. Use this skill when the user needs rendered PNG/SVG output such as visual ERDs (Entity Relationship Diagrams), UI mockups, wireframes, or architecture illustrations. TRIGGER when: user asks for PNG/SVG output, UI mockups, wireframes, visual ERDs, or says "generate image" / "create mockup". DO NOT TRIGGER when: text-based Mermaid diagrams (use external-diagram-mermaid-generate), or non-visual documentation tasks.
Run an investment idea through Charlie Munger's latticework of mental models and surface which models "pop" — which fire for the idea, which fire against it, and what an inversion pass reveals. A decision-support thinking tool, not financial advice. Use when the user invokes /munger, says "run this idea through mental models", "munger check this", "what mental models apply to this investment", or pastes an investment thesis / ticker / business and asks which models fire.
Choose the right MoE token dispatcher (`alltoall`, DeepEP, or HybridEP) for the hardware, EP degree, and optimization stage. Summarizes patterns from DSV3, Qwen3, Qwen3-Next, and VLM bring-up work.
Generate a source-backed starting `trtllm-serve --config` YAML for basic aggregate single-node PyTorch serving, aligned with checked-in TensorRT-LLM configs and deployment docs. Preserves explicit latency / balanced / throughput objectives. Excludes disaggregated, multi-node, and non-MTP speculative configs.
Serve a quantized or unquantized LLM checkpoint as an OpenAI-compatible API endpoint using vLLM, SGLang, or TRT-LLM. Use when user says "deploy model", "serve model", "start vLLM server", "launch SGLang", "TRT-LLM deploy", "AutoDeploy", "benchmark throughput", "serve checkpoint", or needs an inference endpoint from a HuggingFace or ModelOpt-quantized checkpoint. Do NOT use for quantizing models (use ptq) or evaluating accuracy (use evaluation).
Help the user define a concrete, measurable goal before starting work, especially when they ask to use the goal tool, create a goal, set an objective, clarify success criteria, or turn a fuzzy intention into a quantitative outcome. Use this skill for goal creation and goal refinement only; it does not manage durable snapshots, decision logs, or long-running execution artifacts.
Practical guidance for training MoE VLMs in Megatron Bridge. Compares FSDP and 3D-parallel approaches, using rounded lessons from Qwen3-VL, Qwen3-Next, and other multimodal experiments.
Onboard 1-node GitHub MR functional tests for GB200 from existing mr-scoped 2-node tests.