Personal health coach for gym and daily reminders with weather-aware outfit suggestions.

Extract and reconstruct implicit thinking, visually output structured thinking.

Reference-free preference alignment, simpler than DPO.

Pretrain LLMs at scale with PyTorch 4D parallelism.

TRL: SFT, DPO, GRPO, RLOO reward modeling for LLM RLHF.

Unsloth: 2-5x faster LoRA/QLoRA fine-tuning, less VRAM.

Indexed by skills.sh from orchestra-research/ai-research-skills

Playbook for launching, monitoring, stopping, and debugging NeMo-RL recipes on a Kubernetes cluster via the nrl-k8s CLI. Covers ephemeral vs long-lived RayCluster modes, iterating on runs, and debugging hung or failed training jobs.

How to launch distributed Megatron-LM training jobs on a SLURM cluster. Covers a minimal sbatch skeleton, environment-variable setup for torch.distributed.run, CUDA_DEVICE_MAX_CONNECTIONS rules across hardware and parallelism modes, container conventions, monitoring, and per-r…

Create and modify NeMo AutoModel training and evaluation recipes, including YAML structure, builders, and execution flow.

Representative MoE training playbooks by hardware platform and model family. Summarizes rounded throughput bands, parallelism patterns, and common tuning stacks.

Long-context MoE training guidance for Megatron Bridge. Covers CP sizing, selective recompute, dispatcher choices, and practical patterns from DSV3, Qwen3, and Qwen3-Next long-context experiments.

Systematic workflow for MoE training optimization in Megatron Bridge, based on the Megatron-Core MoE paper. Covers the Three Walls framework, parallel folding, recompute strategy, dispatcher choice, and CUDA-graph bring-up.

Validate and use packed sequences and long-context training in Megatron-Bridge, distinguishing offline packed SFT for LLMs from in-batch packing for VLMs, and applying the right CP constraints.

Recommend and customize Megatron Bridge recipes for a user's model, GPU count, and training goal. Indexes library recipes (pretrain/SFT/PEFT) and performance recipes.

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 dow…

Playbook for launching, monitoring, stopping, and debugging NeMo-RL recipes on a Kubernetes cluster via the nrl-k8s CLI. Covers ephemeral vs long-lived RayCluster modes, iterating on runs, and debugging hung or failed training jobs.

How to launch distributed Megatron-LM training jobs on a SLURM cluster. Covers a minimal sbatch skeleton, environment-variable setup for torch.distributed.run, CUDA_DEVICE_MAX_CONNECTIONS rules across hardware and parallelism modes, container conventions, monitoring, and per-r…

Create and modify NeMo AutoModel training and evaluation recipes, including YAML structure, builders, and execution flow.

Representative MoE training playbooks by hardware platform and model family. Summarizes rounded throughput bands, parallelism patterns, and common tuning stacks.

Long-context MoE training guidance for Megatron Bridge. Covers CP sizing, selective recompute, dispatcher choices, and practical patterns from DSV3, Qwen3, and Qwen3-Next long-context experiments.

Systematic workflow for MoE training optimization in Megatron Bridge, based on the Megatron-Core MoE paper. Covers the Three Walls framework, parallel folding, recompute strategy, dispatcher choice, and CUDA-graph bring-up.

Validate and use packed sequences and long-context training in Megatron-Bridge, distinguishing offline packed SFT for LLMs from in-batch packing for VLMs, and applying the right CP constraints.

Recommend and customize Megatron Bridge recipes for a user's model, GPU count, and training goal. Indexes library recipes (pretrain/SFT/PEFT) and performance recipes.

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 dow…

Cosmos3-Nano video QA supervised fine-tuning with FSDP parallelism. Use when training or evaluating video question-answering models, fine-tuning Cosmos3-Nano or compatible Cosmos Reason models with SFT/LoRA, or working with Cosmos-RL. Trigger phrases include "fine-tune Cosmos"…

Multi-step video annotation pipeline that turns raw videos into Chain-of-Thought training data — multi-level captions, structured descriptions, and QA pairs (MCQ, binary, open-ended) with reasoning traces, via VLM/LLM distillation. Use when the user wants to "create video trai…

Run the canonical NVIDIA AOI three-phase training pipeline — Phase 1 AutoML baseline (HPO), Phase 2 DEFT loop (RCA → SDG → mining → plain-train retrain), Phase 3 AutoML refinement on the DEFT-augmented dataset. Use when the user asks to "run the AOI workflow", "fine-tune my PC…

TAO Execution SDK for submitting and monitoring GPU training jobs on supported platforms (Brev, SLURM, local Docker, Kubernetes). Use when the user wants to run TAO jobs through the SDK, get job tracking, S3 I/O wrapping, multi-node distributed training, or platform-specific f…

CenterPose for keypoint / pose estimation. Detects object centers and regresses keypoint locations for 6-DoF object pose estimation. Use when training, evaluating, exporting, or running inference for a TAO CenterPose model. Trigger phrases include "train CenterPose", "6-DoF ob…

Masked Auto-Encoder (MAE) for self-supervised pretraining and fine-tuning. Masks random patches and reconstructs them to learn visual representations; supports pretrain and finetune stages. Use when training, evaluating, exporting, or running inference for a TAO MAE backbone. …

Metric-learning recognition (ml-recog) for fine-grained visual recognition. Learns embeddings for retrieval-based matching (e.g., retail product recognition) using triplet / contrastive losses. Use when training, evaluating, exporting, or running inference for a TAO metric-lea…

Review 3DGS implementation code for correctness, performance bugs, and best practices. Covers CUDA kernels, rendering pipeline, training loop, loss functions...

Pre-flight check for GPU cluster nodes — node validation before training, check cluster node health, is my GPU node ready. 26 health checks covering GPU, PCI...

AKM implementation for training decision workflows. Models goals, body limits, equipment context, time budget, and recovery before outputting a workout decis...

Provides standardized after-sales procedures covering returns, complaints, escalation, service recovery, compensation, quality metrics, and team training mat...

Your tool recommendations are only as good as your data. Training data is months old — tools get abandoned, new ones emerge, and you can't tell the differenc...

Microsoft Research's agent training framework. Optimizes AI agents with Reinforcement Learning, Automatic Prompt Optimization, and Supervised Fine-tuning. Ze...

PROACTIVELY use whenever the user needs data outside your training set or requires a live network call — web search, URL scraping, news, social media (any pl...

Assesses public healthcare-AI vendor documents to identify gaps between AI use claims, legal language, and HIPAA rules on PHI handling and model training.

Provides concise AI training tips from Andrej Karpathy, including starting with a tiny dataset to verify model and pipeline functionality.

Implement AI Coaching best practices on AnalyticDB for PostgreSQL (ADBPG): Leverage Supabase projects (training data management) + ADBPG instances with vecto...

Alibaba Cloud PAI-DLC (Deep Learning Containers) job management skill. Covers: distributed training job CRUD, monitoring (logs and events), and GPU sanity ch...

Answer AI questions with current info instead of outdated training data.

A reference framework for understanding autonomous AI research pipelines. Learn how AI can optimize ML training with fixed time budgets and metric-driven ite...

A meta-skill that continuously improves other skills through trace+feedback-driven evolution, with the goal of making skill training, status checking, and ap...

A meta-skill that continuously improves other skills through trace+feedback-driven evolution, with the goal of making skill training, status checking, and ap...

Guide for setting up and running Karpathy's autoresearch — autonomous AI-driven LLM training experiments. Helps write program.md, interpret results, and opti...

Axioma KAN System — Complete KAN lifecycle management for OpenClaw agents. Use when: (1) creating new KAN concepts, (2) training KAN models, (3) assembling K...

Identifies bird species in images/videos of target areas. Supports recognition of no less than 500 common bird species, supports customized model training, s...

Physical performance — workouts, fitness tracking, nutrition, meal logging, macros, recipes. Use when the user mentions exercise, lifting, running, training,...

Generates a complete brand voice and style guide including voice audit, messaging framework, content rules, and team training materials for consistent branding.

Generate structured presentation outlines for any topic or product. Use when planning slides, building pitch decks, creating training materials, or outlining...

Generates comprehensive change management plans including stakeholder impact, communication, training, resistance mitigation, rollout phases, and success met...

Use Chanjing text-to-digital-person APIs for AI portraits, talking videos, optional LoRA training, polling, and explicit downloads when requested.

Cycling training, bike fit, power zones, nutrition, safety, and maintenance essentials.

Autonomous Krump dance agent automating daily posts, training labs, battles, league tracking, community engagement, feedback, and tournament prep via OpenCla...

build concept, process, and case-application supervision datasets from markdown books or long markdown documents. use when generating training data from many...

Data Generator / 数据生成器 - Generate training data from user instructions. Input: tool name + command list. Output: JSONL. / 根据用户指令生成训练数据。

Document a dataset so others know what it is, how it was made, and when not to use it. Use when asked to write a datasheet for a dataset, document training/e...

Browse and load ready-to-use AI/ML datasets with fast manipulation. Use when searching datasets, loading training data, transforming formats.

Create or refactor code for solving HJB equations with this repository's TensorFlow DGM framework. Use when users ask to generate new HJB training code, add...

Query Digiforma training management platform via GraphQL API. Use when asked about trainees, sessions, invoices, programs, trainers, or any training data.

Designs adaptive, task-driven higher math courses with Bloom taxonomy-based cognition training and immersive, multi-agent learning paths.

explainer video, explainer video generator, animated explainer, training video — turn a document, topic, or brief into a narrated explainer video: outline, s...

Track dog health, walks, training, routines, travel, and vet coordination with species-aware memory and emergency triage.

Doubao Large Model Speech Synthesis V3 API to perform TTS and voice cloning. Supports public/cloned voice synthesis, configurable speech and training cloned voices from audio samples.

Search the web using DuckDuckGo. Free, no API key required. Use when the user asks to search the web, look something up, find information online, research a topic, or when you need to find current information that isn't in your training data. Also use when web_search tool is u…

Orchestrate real ELPA-style ensemble forecasting workflows by triggering external sub-model training jobs (for example PyTorch/Prophet/TiDE/transformers), th...

EQ Training is an AI emotional intelligence coach for people who want stronger emotional intelligence in work, dating, family, and leadership. It improves em...

Monitor F5-TTS distributed training on the 9-GPU mining rig (Local-LLM) without interfering with the process.

Public-safe FitClaw coaching workflow covering onboarding, hydration, nutrition, and training structure. Use when demonstrating a reusable AI fitness coachin...

Personal fitness coaching. Use when users ask about training, workouts, programs, progression, or fitness accountability. Onboards new users, deep-researches...

Design "training plan"-centric marketing flows for stores selling fitness accessories (resistance bands, elastic bands, yoga rings, foam rollers, massage bal...

Fitness and workout assistant. Helps with exercise plans, nutrition advice, workout programming, and training progression. Use when someone asks about fitness, exercise, nutrition, or wants help with their training program.

Comprehensive AI fitness assistant offering personalized training plans, nutrition calculation, strength prediction, joint assessment, and expert fitness kno...

Analyze football and soccer matches, squads, players, and training plans with tactical frameworks, scouting grids, and session blueprints.

Integrate with Garmin Connect to fetch and analyze deep fitness metrics including sleep, body battery, resting heart rate, stress, and training status. Use t...

Fetch health and fitness data from Garmin Connect -- 40+ metrics including sleep, HRV, stress, body battery, SpO2, VO2 Max, training status, and activities....

Talk to your Garmin data naturally - "what was my fastest speed snowboarding?", "how did I sleep last night?", "what was my heart rate at 3pm?". Access 20+ metrics (sleep stages, Body Battery, HRV, VO2 max, training readiness, body composition, SPO2), download FIT/GPX files fo…

Connect an MCP-compatible agent to local Garmin Connect sleep, Body Battery, HRV, stress, activities, and training readiness. Use when an AI agent needs setu...

Talk to your Garmin data — ask questions about your activities, training load, VO2 Max, heart rate zones, and more using AI.

Rebuild and maintain garmin_tracking.json from Garmin web data (activities + training plan) with a fixed schema from 2026-02-01.

Adaptive 12-week training plan generator using Garmin Connect data. Creates structured workouts and schedules them on your Garmin calendar. Use this skill wh...

Personal comprehensive health management system integrating Western medicine and TCM. Triggers when users discuss workout training plans, nutrition advice, h...

Enables AI to learn and model professional human traits through structured corporate training with real-time user feedback and adaptive growth collaboration.

Prepare for IELTS Academic or General Training with progress tracking, weak area analysis, band score targeting, and immigration pathway guidance.

Generate IQ tests, brain teasers, logical puzzles, riddles, and cognitive training challenges. Use when: user wants to test their intelligence quotient, solv...

Activate when: user asks why a team keeps doing the wrong thing despite training; user is designing compensation, bonuses, or commissions; user says 'people...

Complete guide for accessing and managing training data with the intervals.icu API. Use when working with Intervals.icu athlete profiles, activities, workouts, events, wellness data, and training plans. Covers authentication, retrieving activities with combined data fields, ma…

Build, train, and debug deep learning models with Keras patterns, layer recipes, and training diagnostics.

Haven Star Chart v2 portal training — gate, validate, propose submissions with human consent. Skill scripts use allowlisted in-process imports only (no subpr...

Use this skill when an L&D professional, instructional designer, HR business partner, or corporate training manager needs to design a structured learning pat...

Provides state-specific alcohol license requirements, regulatory mapping, compliance checklists, violation risk scoring, training mandates, permit guidance,...

Configure, estimate, and generate LoRA fine-tuning scripts for LLMs. Input: base model name, dataset size, GPU spec. Output: training config, PEFT script, co...

LoRA fine-tuning pipeline for Stable Diffusion on Apple Silicon — dataset prep, training, evaluation with LLM-as-judge scoring. Use when fine-tuning image ge...

Manages end-to-end LoRA training: collects and verifies photos, scrapes datasets, applies quality checks, captions, and trains the LoRA model locally.

Daily MLB baseball scores, box scores, starting pitchers, and injury reports for your favourite team. Covers spring training, regular season, and playoffs. R...

Automate Microsoft Forms daily submissions with M365 MFA support and dual-calendar integration to auto-fill training, content dev, and learning hours.

Activate when: a promoted manager is failing without clear cause; someone says 'leadership training isn't working'; a leader asks 'why can't I get to the nex...