MLOps
模型部署、评估与运维
punderstruck
alexgreensh
Pun expert and comedy brainstorming partner. Discovers wordplay through Datamuse phonetic analysis and structured comedy theory. Use when asking for puns, dad jokes, wordplay, brainstorming creative angles, shower thoughts, roasts, jargon translations, quote remixes, or pun compositions of any length.
replicate-cli
VladimirBrejcha
This skill provides comprehensive guidance for using the Replicate CLI to run AI models, create predictions, manage deployments, and fine-tune models. Use this skill when the user wants to interact with Replicate's AI model platform via command line, including running image generation models, language models, or any ML model hosted on Replicate. This skill should be used when users ask about running models on Replicate, creating predictions, managing deployments, fine-tuning models, or working with the Replicate API through the CLI.
affiliate-onboarding
dmend3z
This skill guides users in setting up a robust affiliate onboarding system. It covers welcome sequences, swipe file creation, training curriculum, compliance guidelines, resource library setup, first promotion guidance, and communication templates to transform new affiliates into high-performing partners.
affiliate-motivation-and-training
dmend3z
This skill equips the agent to motivate and train affiliate partners effectively. It should be used for tasks like creating weekly updates, sharing success stories, developing training modules, writing personal outreach scripts, planning reactivation campaigns, and designing performance coaching and incentive programs.
data-pipeline-engineering
lv416e
"Use when building or modifying any data pipeline, before writing transformation logic - idempotent-first approach covering schema design, quality checks, incremental loads, CDC, and observability that ensures every step is repeatable and verifiable データパイプラインの構築や変更時、変換ロジックを書く前に使用 - スキーマ設計、品質チェック、増分ロード、CDC、オブザーバビリティを網羅する冪等性ファースト手法により、全ステップの再実行可能性と検証可能性を保証"
setup
mickeyyaya
Use when the user runs /evo:setup (or /evo:setup), asks to configure evolve-loop, onboard, pick per-phase models, or learn how the pipeline works. Auto-detects available LLM CLIs/subscriptions, explains the pipeline concisely, then presents THREE ready-made config presets (Recommended/Economy/Max-quality) the Go binary computes deterministically from the public profiles — the user makes ONE choice and the binary writes per-phase pins to .evolve/policy.json. Runs once on first launch (the loop nudges) and is re-runnable anytime.
model-hierarchy
zscole
Cost-optimize AI agent operations by routing tasks to appropriate models based on complexity. Use this skill when: (1) deciding which model to use for a task, (2) spawning sub-agents, (3) considering cost efficiency, (4) the current model feels like overkill for the task. Triggers: "model routing", "cost optimization", "which model", "too expensive", "spawn agent".
python-notebooks-async
ahgraber
"Use when writing or reviewing asyncio code in Jupyter notebooks or '#%%' cell workflows — structuring event-loop ownership, orchestrating async tasks, or choosing compatibility strategies. Also use when hitting RuntimeError: This event loop is already running, asyncio.run() failures in cells, or tasks silently never completing."
audit-context-building
lv416e
Enables ultra-granular, line-by-line code analysis to build deep architectural context before vulnerability or bug finding.
phase-create
mickeyyaya
Use when the user (or the advisor) wants to design and register a NEW optional pipeline phase conversationally — "add a phase that does X". Interviews for goal-type, trigger signal, report sections, and verdict; synthesizes a phase.json + persona; registers them via evolve phases create; self-corrects from the command's JSON envelope. The Go binary is the single enforcement point, so this flow works from any LLM CLI.
cheap-model-testing
almeidamarcell
When working on any application that integrates with LLMs or pay-per-usage APIs, always use the cheapest available model during development and testing. Remind to upgrade to a production model before deployment.
hughes-property-based-testing
copyleftdev
Test software in the style of John Hughes, inventor of QuickCheck and property-based testing. Emphasizes specifying properties that should hold for all inputs, generating random test cases, and shrinking failures to minimal examples. Use when testing algorithms, data structures, parsers, serializers, or any code with clear invariants.
adk-skill
miticojo
"Build single-agent and multi-agent systems using Google's Agent Development Kit (ADK) in Python, Java, Go, or TypeScript. Use when user says 'build an agent with ADK', 'create a Gemini agent', 'multi-agent pipeline', 'agent orchestration with Google', or mentions ADK, google-adk, google agent development kit, sequential/parallel/loop agents, agent tools, callbacks, state management, agent testing, or agent deployment with Gemini. Do NOT use for LangChain, CrewAI, AutoGen, or non-ADK agent frameworks."
feature-pipeline
notedit
Execute implementation tasks from design documents using markdown checkboxes. Use when (1) implementing features from feature-analyzer output, (2) resuming interrupted work, (3) batch executing tasks. Triggers on 'start implementation', 'run tasks', 'resume'.
training-honest-models
omar320250068-prog
Trains a first model the honest way — method chosen to fit the question, compared against the baseline on the same split and metric, errors read before scores are believed. Use when moving from a rule baseline to a learned model, or when reviewing a model that reports only a single score.
framing-ml-problems
omar320250068-prog
Frames a data/ML problem before any modeling — the decision, the action, the cost of a wrong call, task type, target, and success metric. Use when picking a project direction, writing a research question, or mapping a problem onto an ML task type.
mlops-workflows
manutej
Comprehensive MLOps workflows for the complete ML lifecycle - experiment tracking, model registry, deployment patterns, monitoring, A/B testing, and production best practices with MLflow
building-baselines
omar320250068-prog
Builds the transparent rule-based baseline every model must beat — a hand-written score with reason codes, ranked output, and precision@K evaluation. Use before training any model, or when someone reports model results with nothing to compare against.
hunting-leakage-and-validating
omar320250068-prog
Finds label leakage and designs honest validation — leakage taxonomy, grouped and time-aware splits, base rates, and the attack-your-own-model checklist. Use before trusting any metric, when a score looks too good, or when features and labels share time windows or origins.
ollama-optimizer
Montimage
Optimize Ollama configuration for maximum performance on the current machine. Use when asked to "optimize Ollama", "configure Ollama", "speed up Ollama", "tune LLM performance", "setup local LLM", "fix Ollama performance", "Ollama running slow", or when users want to maximize inference speed, reduce memory usage, or select appropriate models for their hardware. Analyzes system hardware (GPU, RAM, CPU) and provides tailored recommendations.
eol-internal-enablement
TurboSnails
"Build the support FAQ, sales talking points, and objection handling teams need before an EOL announcement. Use when customer-facing teams must be ready before customers hear."
fal-serverless-guide
JosiahSiegel
Complete fal.ai serverless deployment system. PROACTIVELY activate for: (1) Creating fal.App class, (2) GPU machine selection (T4/A10G/A100/H100), (3) setup() for model loading, (4) @fal.endpoint decorators, (5) Persistent volumes for weights, (6) Secrets management, (7) Scaling configuration (min/max concurrency), (8) Multi-GPU deployment, (9) fal deploy commands, (10) Local development with fal run. Provides: App structure, Dockerfile patterns, deployment commands, scaling config. Ensures production-ready serverless ML deployment.
fal-optimization
JosiahSiegel
Complete fal.ai optimization system. PROACTIVELY activate for: (1) Queue vs run performance, (2) Parallel request batching, (3) Streaming for real-time UI, (4) WebSocket for interactive apps, (5) Model cost comparison, (6) Image size optimization, (7) Inference step tuning, (8) Webhook vs polling, (9) Result caching by seed, (10) Serverless scaling config. Provides: Parallel patterns, cost strategies, caching examples, monitoring setup. Ensures optimal performance and cost-effective usage.
fal-model-guide
JosiahSiegel
Complete fal.ai model selection system. PROACTIVELY activate for: (1) Choosing image generation models (FLUX, SDXL), (2) Choosing video models (Kling, Sora, LTX), (3) Choosing audio models (Whisper, ElevenLabs), (4) Model quality vs speed comparison, (5) Cost optimization by model tier, (6) 3D generation models, (7) Model-specific parameters, (8) Development vs production model selection. Provides: Model comparison tables, decision trees, pricing tiers, performance benchmarks. Ensures optimal model selection for quality, speed, and cost.