Skills
98,639 packaged capabilities and playbooks for agents. Install any of them with one command.
Development
Discover which models, recipes, and tasks the getitune library (the Geti training library) supports before training. Use when a user asks what models are available, how to list recipes, how to filter by task or name pattern, how `list_models(...)` and `getitune find` behave, or how to resolve the "model name matches multiple tasks" error. Covers classification, detection, instance/semantic segmentation, and keypoint detection recipes.
Export a trained getitune model (the Geti training library) to a deployable format. Use when a user wants to run `engine.export(...)` or `getitune export`, choose between OpenVINO IR and ONNX, set FP32 vs FP16 precision with `ExportFormat` / `Precision`, or understand where exported artifacts are written and how they load back for inference. Covers the export/load contract between training and OpenVINO/ONNX inference.
Optimize an exported getitune model (the Geti training library) with post-training quantization. Use when a user wants to run `OVEngine.optimize()` / `engine.optimize()` to produce an INT8 model via NNCF, understands calibration-set requirements, or needs to re-validate and run inference with a quantized model versus the original FP32/FP16 model. Covers OpenVINO NNCF post-training quantization and the accuracy/size trade-off.
Run inference and evaluation with a getitune model (the Geti training library). Use when a user wants to call `engine.predict()` / `engine.test()` or `getitune predict` / `getitune test`, run inference with a PyTorch checkpoint versus an exported OpenVINO IR (`.xml`) or ONNX (`.onnx`) model, or understand how `OVEngine` loads deployed models via ModelAPI. Covers PyTorch, OpenVINO, and ONNX inference backends.
Train a computer-vision model with the getitune library (the Geti training library) using its Python API or CLI. Use when a user wants to train, fine-tune, or evaluate a model with `create_engine(...)` and `engine.train()/engine.test()`, run `getitune train`/`getitune test`, pick or override a recipe under `getitune.recipe.<task>`, choose a device (cpu/gpu/xpu/cuda), warm-start from a checkpoint, or debug a training run. Covers classification, detection, instance/semantic segmentation, and keypoint detection.
This skill should be used when working with GitHub CLI (gh) for pull requests, issues, releases, and GitHub automation. Use when users mention gh commands, GitHub workflows, PR operations, issue management, or GitHub API access. Essential for understanding gh's mental model, command structure, and integration with git workflows.
Enter a project idea or GitHub link to automatically search for related open-source projects and generate a structured analysis report (tech stack/strengths and weaknesses/score), and download the code packages of the top 3 highest-scoring projects. Supports both intent-based search and direct-link analysis modes.
Control Herdr, a terminal multiplexer for coding agents. Use Herdr to inspect or control panes, tabs, workspaces, terminals, run commands, and or starting and monitoring background processes like dev servers. Use for subagent when the user or another skills explicitly asks for or requies subagents. Requires HERDR_ENV=1.
Use when the user asks to “create a hook,” “add a PreToolUse/PostToolUse/Stop hook,” “validate tool use,” “implement prompt-based hooks,” “use ${CLAUDE_PLUGIN_ROOT},” “set up event-driven automation,” “block dangerous commands,” or mentions hook events (PreToolUse, PostToolUse, Stop, SubagentStop, SessionStart, SessionEnd, UserPromptSubmit, PreCompact, Notification). Provides comprehensive guidance for creating and implementing Claude Code plugin hooks, focusing on the advanced prompt-based hooks API
Use when the user says 'convert this html to breakdance' or 'paste html into breakdance', or has a Webflow, Framer, CodePen, or old static export to bring into a Breakdance site. Converts raw HTML and CSS into native EssentialElements nodes in _breakdance_data, mapping colors, type, and spacing to Breakdance Variables, and refusing element types Breakdance would skip-render.
Research-backed UX best practices for Shopify Hydrogen storefronts, based on Baymard Institute (200,000+ hours of usability research) and Nielsen Norman Group studies. Use this skill WHENEVER the user is building, reviewing, or improving any part of a Shopify Hydrogen e-commerce site — including homepage, collection/listing pages, product pages, cart, checkout, search, filters, mobile UX, accessibility, or trust/credibility concerns. Also use when the user asks for code review on Hydrogen components, wants to scaffold a new route or component, or mentions conversion, cart abandonment, or shopping experience issues. Trigger even if the user doesn't explicitly say "UX" — any Hydrogen storefront work benefits from these guidelines.
Elite website image-to-code skill for Codex. For visually important web tasks, it must first generate the design image(s) itself, deeply analyze them, then implement the website to match them as closely as possible. In Codex, it must prefer large, readable, section-specific images instead of tiny compressed boards, generate fresh standalone images for sections or detail views instead of cropping old ones, avoid lazy under-generation, avoid cards-inside-cards-inside-cards UI, and keep the hero clean, spacious, readable, and visible on a small laptop.
Optional meta-skill for improving `chrislacey89/skills` from real-world friction or breakdowns discovered while using the pipeline in another repo. Use when the main lesson is about the pipeline itself, not the downstream project. Grounds proposals in established software-engineering guidance from `/library`. Produces a GitHub issue in `chrislacey89/skills` and only moves to implementation after review.
Add PostHog metric capture to track counters, gauges, and histograms. Use after implementing features or reviewing PRs to ensure key operations are measured, or when asked to add metrics, counters, or track a value over time. Also handles mirroring existing Prometheus/StatsD/OpenTelemetry metrics into PostHog and initial OTLP exporter setup.
Design, implement, review, and troubleshoot Pakistan 1LINK 1BILL integrations for bill inquiry, bill payment, fixed invoice or voucher collections, variable or top-up payments, credit-card bill payments, packages, payment-status inquiry, voids, PSID or consumer-number flows, reconciliation, and certification. Use when a bank, fintech, government entity, educational institution, utility, biller, or Pakistani business needs to add or repair a 1BILL collection workflow.
Design, implement, review, migrate, and troubleshoot secure JazzCash merchant payment integrations for Pakistani websites, apps, and commerce platforms. Cover HTTP POST redirection, approved APIs or hosted checkout, JazzCash Mobile Wallet, voucher payments, cards when enabled, secure hashes, callbacks, status checks, refunds, settlement, reconciliation, onboarding, testing, and production rollout. Use when adding JazzCash, repairing an existing JazzCash gateway, or upgrading an outdated merchant integration.
Design, implement, review, and troubleshoot secure Stripe integrations for one-time payments, Checkout, Payment Element, subscriptions, invoices, saved payment methods, refunds, disputes, webhooks, customer management, Stripe Tax, and Connect platforms or marketplaces. Use when adding Stripe to a web or mobile backend, building SaaS billing, accepting payments, creating connected accounts, or preparing an existing Stripe integration for production.
JSON processing with jq for filtering, transforming, aggregating, and reshaping data from files, APIs, and command output on the command line. Use when processing JSON via command-line filters. Don't use for CSV/TSV (use xsv), YAML (use yq), or analytical queries requiring SQL (use duckdb).
External-research routing for confirmed code problems: triage whether the root cause is internal / external / hybrid, run a known-issue quick search before deep root-causing (platform silent failures, nested host runtimes, no code-level suspects), and evaluate industry-wide hard limits. Delegates all WebSearch discipline to effective-web-research. Referenced by PDCA hosts via frontmatter dependencies; load during the technical-analysis stage. Do NOT use for competitor product-behavior, interaction-boundary, or mechanism observation (competitive-research).
Provides comprehensive guidance for Koa.js framework including middleware, context, async/await patterns, and application structure. Use when the user asks about Koa, needs to create Koa applications, implement middleware, or build Node.js web applications.
Best practices for Kuroco API design and content management. Covers endpoint design, authentication (Cookie, dynamic token, static access token), CORS, rate limiting, caching, error handling (401/403/429), content CRUD via the Topics API (list / details / insert / update / delete / bulk_upsert), filter queries, pagination, multilingual support (langs_open_flg), extended fields (ext_col), file upload and CSV import, and EC points (ECPoint). Use for implementing API calls, login API and token authentication, resolving authentication and permission errors, and questions about fetching, creating, updating, deleting, and filtering content.
Java logging best practices with SLF4J, structured logging (JSON), and MDC for request tracing. Includes AI-friendly log formats for Claude Code debugging. Use when the user asks about logging, debugging application flow, or analyzing logs.
Use when adding or fixing Luau type annotations - luau-lsp/luau-analyze errors, --!strict compliance, "Type X could not be converted into Y", generics, union types, type narrowing/refinements, casts (::), typed metatables and OOP, .d.luau definitions, or .luaurc setup. Triggers include "type error", "strict mode", "luau-lsp", "type mismatch", "export type".
Review pull request feedback with discerning judgment. The lynx scans comments, identifies legitimate concerns, addresses minor issues directly, and plans responses to significant feedback. Use when responding to PR reviews and addressing reviewer concerns.
MapStruct Java mapping framework. Covers basic mappers, update mappings, nested objects, collections, custom methods, and Lombok integration. Based on castellino and gestionale-presenze production patterns. USE WHEN: user mentions "mapstruct", "@Mapper", "@Mapping", asks about "DTO mapping", "entity conversion", "@MappingTarget", "nested mapping" DO NOT USE FOR: Java language - use `java` skill instead DO NOT USE FOR: Lombok - use `lombok` skill instead DO NOT USE FOR: JPA entities - use JPA-specific skills
Material Design 3 (Material You) design system knowledge for modern web and Angular applications. Use when implementing Material Design 3 theming, components, typography, color systems, dynamic color, accessibility patterns, or migrating from Material Design 2. Covers design tokens, theming APIs, and Material You principles.
Before fixing a slow/stale/timeout alert, measure the step yourself. Kill the theory with a stopwatch, not a code change. Measure-first ops triage for temporal alerts (stale, timeout, freshness, wedged, N hours behind) from gbrain doctor, autopilot, sync, and cron monitors — runs BEFORE any timeout raise, threshold change, or pipeline rewrite.
MERN stack patterns including React with Vite, Express middleware, MongoDB schemas, API Gateway architecture, session management, error handling, and testing strategies. Activate for MERN development, microservices architecture, and full-stack JavaScript applications.
Provides guidance for enterprise-grade RL training using miles, a production-ready fork of slime. Use when training large MoE models with FP8/INT4, needing train-inference alignment, or requiring speculative RL for maximum throughput.
Wizard that scaffolds a complete Mercado Pago integration from official and bundled sources, using MCP only for selected account actions or documentation gaps. Use whenever the developer wants to add or migrate a Mercado Pago payment flow.
Designs, builds, debugs, and documents n8n workflows and AI agent automations. Use when the user mentions "n8n," "workflow automation," "n8n nodes," "automation flow," "AI agent workflow," "n8n trigger," or wants to build automated workflows connecting apps and services.
(NS) Generate structured requirements.md from scope descriptions, stack context, and design briefs for a product version. Use when starting a new version, the user asks for requirements, RF/features document, or before task generation when requirements.md is missing — even if they only describe the feature in chat. Follow SDD planning gates in ns-harness. Do NOT generate task files (use ns-sdd-task-generator).
This skill should be used when managing objectives - coherent goals that require 2+ plans/PRs to complete. Use when users want to create multi-plan tracking documents, log progress across related PRs, capture lessons learned, or coordinate work that spans multiple erk-prs. Objectives are human-first coordination documents, not executable plans.
GPU-accelerates scientific Python on NVIDIA hardware and verifies that the result is correct and faster. Use for CUDA/GPU optimization; CPU-bound NumPy, SciPy, pandas, scikit-learn, NetworkX, scikit-image, vector-search, image-processing, graph, simulation, or file-I/O workloads; CuPy, cuDF, cuML, cuGraph, cuVS, cuCIM, KvikIO, Warp, Newton, Numba-CUDA, or RAFT questions; and profiling, memory-transfer, kernel, or multi-GPU bottlenecks. Also use when large data-parallel Python code is slow and GPU acceleration is a plausible option, even if the user does not name CUDA.
Enables Flash Attention for transformer models using PyTorch native scaled_dot_product_attention (PyTorch 2.2+) or the flash-attn library, including multi-query attention, sliding window attention, and FP8 on H100 (FlashAttention-3). Covers profiling speedup, checking accuracy against a baseline, and troubleshooting install and GPU support errors. Use when training or running transformers on long sequences (over 512 tokens), when standard attention runs out of GPU memory, when attention is the inference bottleneck, when switching a PyTorch model to the flash backend, or when tuning attention on H100 GPUs. Not for CPU inference, V100 GPUs, or sequences under 256 tokens; consider xFormers for other attention variants.
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.