productivity▌

Token optimization strategies for cost-effective Claude Code usage across all projects. \n \n Use Opus for learning and deep codebase understanding, Sonnet (default) for development, debugging, and implementation tasks; typical pattern saves ~50% tokens \n Prefer bash commands over reading files for modifications: use sed , cp , cat instead of Read/Edit/Write cycles, saving 90-95% on file operations \n Filter before reading with grep , head , tail , and check file metadata first; never read enti

Routing orchestrator for game development projects with platform and dimension-specific sub-skills. \n \n Routes to specialized sub-skills based on target platform (web, mobile, PC, VR/AR) and game dimension (2D or 3D) \n Covers core principles applicable across all platforms: game loop architecture, fixed timestep physics, input abstraction, and performance budgeting for 60 FPS \n Includes pattern selection matrix (State Machine, ECS, Behavior Trees, Object Pooling) with decision rules for when

Postgres performance optimization rules across 8 priority categories, from query tuning to advanced features. \n \n Organized into 8 rule categories prioritized by impact: query performance and connection management (critical), security and RLS, schema design, concurrency, data access patterns, monitoring, and advanced features \n Each rule includes detailed explanations, incorrect vs. correct SQL examples, EXPLAIN output analysis, and performance metrics to guide optimization decisions \n Cover

Complete FilamentPHP v4 resources with forms, tables, relation managers, and custom pages. \n \n Generates resource classes with form schemas, table configurations, and page routing following official FilamentPHP documentation patterns \n Supports resource variants including soft deletes, view-only pages, and simple modal-based forms via artisan commands \n Includes 20+ form field types (text inputs, selectors, date/time pickers, file uploads, relationship fields) and 5+ table column types with

Orchestrates category-specific CTF skills to solve challenges across pwn, crypto, web, reverse engineering, forensics, OSINT, malware, and misc. \n \n Performs initial reconnaissance by exploring files, analyzing binaries with strings / binwalk / checksec , fetching remote URLs, and connecting to services to understand challenge context \n Categorizes challenges by file type, keywords, and service behavior, then invokes the matching specialized skill ( /ctf-web , /ctf-pwn , /ctf-crypto , etc.) \

Grill me about specs for the workflows I want to build, within this workspace.

Structured commit message format following the Conventional Commits specification. \n \n Defines 12 commit types (feat, fix, refactor, docs, test, chore, ci, build, perf, style, deps, revert) with clear purposes and semantic versioning correlation \n Enforces a required header format with optional scope and body, all lines under 100 characters, subject line capped at 70 characters \n Includes guidelines for imperative present-tense language, body content explaining what and why, and support for

Comprehensive, library-agnostic checklist for reviewing Flutter/Dart applications. These principles apply regardless of which state management solution, routing library, or DI framework is used.

本技能基于 F2 框架实现抖音视频批量下载，提供高效、稳定的批量下载能力。

Automatically analyze Bilibili videos, extract frames, and generate structured documentation or tutorials using AI. \n \n Downloads videos via Bilibili API, extracts frames using FFmpeg at configurable rates (1 fps for short videos, 0.5 fps for medium, 0.2 fps for long), and deduplicates similar adjacent frames \n Batch-analyzes extracted frames in parallel using the Task tool, with dynamic batching strategy based on total frame count (1–5 batches) \n Reorganizes analysis results into polished m