Loading...
Loading...
Found 5,545 Skills
Train reasoning and verifiable-task behavior with GRPO and reinforcement learning from verifiable rewards (RLVR). Use when task success is algorithmically checkable (math, code, tool calls, structured output), when designing GRPO reward functions, or when a GRPO run diverges or reward-hacks.
How to work in a Codebase Wiki project (the `codebase-wiki` starter pack): an agent-authored, source-grounded wiki of the surrounding codebase. Read when the project has a `wiki/` knowledge base with `architecture/`, `modules/`, `flows/`, `concepts/`, and `guides/` sections plus `wiki/OVERVIEW.md`, or when asked to generate or refresh a wiki of this codebase. Carries the per-folder rules and freshness + log discipline, summarizes the audience/depth knobs and source-reference convention, and bundles the full generate/refresh procedure in `references/`. Complements the platform `open-knowledge` skill; does not replace it.
Extract structured resources (icons, metadata, text, forms, videos, social links) from any webpage using playwright-cli. Supports individual collectors via subcommands (icons, metadata, text, forms, videos, socials) or all at once. The icon collector classifies SVGs as icon/logo/image based on size and DOM context, optimizes them for EDS, and outputs to /icons/ for use with decorateIcons(). Use when migrating pages, auditing sites, or extracting assets.
Market data preparation including OHLCV resampling, gap handling, anomaly detection, normalization, and multi-source merging
Complete environment variable management with type-safe validation, Vercel dev workflow, and prebuild validation.
Develops resources for FiveM using the QBCore Framework. Covers resource creation, Core Object usage, Player management, Callbacks, Events, Items, Jobs, Gangs, Database (oxmysql), and best practices. Use when the user works with FiveM, QBCore, Lua scripts for QBCore servers, or mentions `QBCore.Functions`, `GetCoreObject`, `CitizenID`, or any system of the QBCore Framework.
Multi-source search and deduplication layer with intent-aware scoring. Integrates Brave Search (web_search), Exa, Tavily, and Grok to provide high-coverage, high-quality results. Automatically classifies query intent and adjusts search strategy, scoring weights, and result synthesis accordingly. Activated for "deep search", "multi-source search", or when high-quality research is needed.
Alba serializers + Typelizer for type-safe Inertia Rails props with auto-generated TypeScript types. Use when serializing models for Inertia responses, setting up Alba resources, generating TypeScript types from Ruby, or using Inertia prop options (defer, once, merge, scroll) with Alba attributes. Replaces as_json with structured, auto-typed ApplicationResource, page resources, and shared props resources. When active, OVERRIDES the `render inertia: { ... }` pattern from other skills — use convention-based instance variable rendering instead.
Track parcels via the 17TRACK API (local SQLite DB, polling + optional webhook ingestion)
Use when asking about 'purge files', 'storage pressure', 'disk space iOS', 'isExcludedFromBackup', 'URL resource values', 'volumeAvailableCapacity', 'low storage', 'file purging priority', 'cache management' - comprehensive reference for iOS storage management and URL resource value APIs
Guide for theming .NET MAUI apps — light/dark mode via AppThemeBinding, ResourceDictionary theme switching, DynamicResource bindings, system theme detection, and user theme preferences. Use when: "dark mode", "light mode", "theming", "AppThemeBinding", "theme switching", "ResourceDictionary theme", "dynamic resources", "system theme detection", "color scheme", "app theme", "DynamicResource". Do not use for: localization or language switching (see .NET MAUI localization documentation), accessibility visual adjustments (see .NET MAUI accessibility documentation), app icons or splash screens (see .NET MAUI app icons documentation), or Bootstrap-style class theming (see Plugin.Maui.BootstrapTheme NuGet package).
Decide whether to fine-tune at all, and route to the right method (SFT, DPO/ORPO/KTO, GRPO/RLVR, continued pretraining) and base model. Use when starting any fine-tuning effort, when unsure whether RAG or prompting would suffice, or when choosing between preference-optimization and reinforcement methods.