Skills
97,725 packaged capabilities and playbooks for agents. Install any of them with one command.
Development
Optional code-delivery discipline: decide whether this run needs commit + PR/MR, then commit/push via git-commit and create or update the PR/MR. Use after verification (and OpenSpec archive when the host requires it) and before feature-branch-closeout. Do NOT use for protected-branch merge (merge-discipline) or release→main shipping (git-release-finish). Triggers — 「提交并开PR」「提交并开MR」「代码交付」「开PR」「开MR」「delivery-discipline」 / deliver code, commit and open PR, open merge request.
Independently and read-only review a design doc, PRD, spec, or requirements for repository accuracy, material risks, and canonical story readiness. Produces a binary readiness verdict with evidence-backed blocking findings and optional non-blocking suggestions. Use design-doc-review-loop when revisions or feature-delivery readiness are requested.
One-command dev loop boot. Spins up portless (named HTTPS subdomain), emulate (stateful API mocks), the project's dev server, and an agent-browser session, all keyed to the current git branch. Use when starting a feature branch, switching worktrees, or returning to a project after a break. Skips silently with install hints when prerequisite binaries are missing.
Build Solana wallet-connected apps with Phantom Connect SDKs and DFlow spot trading. Use when user asks to connect a Phantom wallet, integrate Phantom in React, React Native, or vanilla JS, sign messages or transactions, build token-gated pages, mint NFTs, accept crypto payments, or swap/stream tokens with DFlow. Covers @phantom/react-sdk, @phantom/react-native-sdk, @phantom/browser-sdk, and DFlow spot trading and market-data streaming. Do NOT use for Ethereum or EVM wallet integrations, or non-DFlow DEX routing.
Disciplined diagnosis loop for hard bugs and performance regressions. Reproduce → minimise → hypothesise → instrument → fix → regression-test. Use when user says "diagnose this" / "debug this", reports a bug, says something is broken/throwing/failing, or describes a performance regression.
Create production-quality Django REST Framework APIs using Clean Architecture and SOLID principles. Covers layered architecture (views, use cases, services, models), query optimization (N+1 prevention), pagination/filtering, JWT authentication, permissions, and production deployment. Use when building new Django APIs, implementing domain-driven design, optimizing queries, or configuring authentication. Applies Python 3.12+ and Django 5+ patterns.
Implement drag and drop using @atlaskit/pragmatic-drag-and-drop. Use when implementing sortable lists, reorderable items, kanban boards, or any drag-drop interactions. Covers draggable setup, drop targets, edge detection, drag previews, and critical state management patterns to avoid performance issues.
Use Electric with Drizzle ORM or Prisma for the write path. Covers getting pg_current_xact_id() from ORM transactions using Drizzle tx.execute(sql) and Prisma $queryRaw, running migrations that preserve REPLICA IDENTITY FULL, and schema management patterns compatible with Electric shapes. Load when using Drizzle or Prisma alongside Electric for writes.
Work an entire epic of tickets unattended — for each ticket: plan it, implement it, PR it into an epic branch, drive CI green, merge; keep the epic branch synced with its target; then verify the assembled epic once, open an epic PR that merges cleanly, and file follow-ups. Works with any tracker: GitHub, Jira, GitLab, or a markdown checklist. Use when the user hands over a batch of related issues and wants them worked through autonomously ("run this epic", "work through these tickets", "do the whole milestone"). For a single ticket where the work deserves an adversarial implement/review loop, use `player-coach` instead.
Master error handling patterns across languages including exceptions, Result types, error propagation, and graceful degradation to build resilient applications. Use when implementing error handling, designing APIs, or improving application reliability.
Takes a GitHub PR URL or the current branch and writes a plain-English explanation of what it does and why, then posts it as a PR comment. Use when asked to explain a PR, summarize a pull request, write a plain-English description of a PR, add a summary comment to a PR, or understand what a PR changes. Trigger when a user says "explain this PR", "summarize this pull request", "what does this PR do", "add a comment explaining the PR", or shares a GitHub PR URL and asks what it does.
Orchestrate product-aware Flutter UI design and redesign from repository inspection through rendered verification. Use for new Flutter screens, multi-screen flows, major visual redesigns, design-led implementation, or ambiguous Flutter UI requests that require routing across audit, design-system, implementation, motion, accessibility, and visual-QA workflows.
Maintains the showcase/flutter_test_app macOS showcase, native intentcall hooks (codegen, app_links invoke), and VM MCP validation. Use when editing macOS Runner, intentcall_codegen.sh, macOS dogfood, Screen Recording capture, or comparing macOS parity to web WebMCP.
A skill covering the layer structure, directory conventions, and dependency direction of a Flutter single-package project. Use it whenever questions arise about where to put a new feature, which directory a new screen, repository, or use case belongs in, "how should I set up the folder structure," "where should this feature live," or "should this go in core." Triggers on phrases like "project structure," "directory structure," "layer separation," "clean architecture," "where does X live," and "adding a new feature."
Covers building, extending, debugging, and maintaining Foundry VTT modules for v14+. This skill applies when scaffolding a new module, writing custom Actor/Item types with TypeDataModel, building ApplicationV2 sheets, dialogs, or detached windows, registering hooks or settings, implementing socket communication, extending the canvas with PIXI.js, working with Scene Levels or Regions, managing compendium packs, using ActiveEffect V2 changes, setting up TypeScript with fvtt-types, configuring Vite/Rollup builds, localizing strings, or migrating modules from v13 to v14. Triggers on: "Foundry module", "FVTT", "FoundryVTT", "foundryvtt module", "ApplicationV2", "TypeDataModel", "actor sheet", "module.json", "fvtt-types", "libWrapper", "socketlib", "Scene Levels", "Regions", "ActiveEffect V2", "messageMode", "detached windows", or any task involving Foundry VTT module development.
Use for frontend architecture boundaries: code ownership, directory roles, file placement, import direction, UI/component extraction, API/data contracts, fetching/query, URL state, mappers/schemas/DTOs, hooks/stores, architecture docs, and tool enforcement. Also use before agent-initiated architecture-enforcement changes in lint, CI, import-boundary checks, or similar automation—even without a user request—and for small UI changes that may leak business rules or external data into lower-level code.
Forensic case investigation with evidence-graded findings, calibrated to the input. Use when the user asks to investigate a bug, trace what caused an incident, walk through unfamiliar code, or build a mental model of a code area before working on it.
Implements any user intent, requirement, story, bug fix or change request by producing clean working code artifacts that follow the project's existing architecture, patterns and conventions. Use when the user wants to build, fix, tweak, refactor, add or modify any code, component or feature
Use when the user asks to rewrite recent non-Conventional commit subjects, fix existing commit messages, or clean up commit history. Trigger on phrases like "rewrite commit history", "fix these commit messages", "커밋 메시지 다시 써줘", "/git-commit-rewrite", "$git-commit-rewrite", or whenever the user shows a non-Conventional `git log` and asks for help. For new commits use `git-commit`; for commit + push use `git-commit-push`.
Use when the user asks to merge the current branch into the dev branch (`dev`, falling back to `develop`) and delete the merged source branch unless the source is protected. Trigger on phrases like "merge to dev", "dev에 머지", "develop에 합쳐줘", "merge this into dev and delete it", "/git-merge-to-dev", or "$git-merge-to-dev". For `main` use `git-merge-to-main`; to bulk-clean already-merged branches use `git-branch-cleanup`.
Use when the user asks to merge the current branch into `main` and delete the merged source branch unless the source is protected. Trigger on phrases like "merge to main", "main에 머지", "main 브랜치에 합쳐줘", "merge this branch into main and delete it", "/git-merge-to-main", or "$git-merge-to-main". Refuses to force-delete or to operate on a dirty working tree. For `dev` use `git-merge-to-dev`; to bulk-clean already-merged branches use `git-branch-cleanup`.
Builds and debugs Godot 4.x moment-to-moment 2D gameplay: CharacterBody2D movement, platformer or top-down physics, TileMapLayer worlds, animation, hitboxes, projectiles, health, combat feedback, cameras, and game feel. Use for player controllers, jumping, collisions, tilemaps, damage, shooting, and 2D polish; use godot-programming-patterns for pooling architecture.
Post-training 4-bit quantization for LLMs with minimal accuracy loss. Use for deploying large models (70B, 405B) on consumer GPUs, when you need 4× memory reduction with <2% perplexity degradation, or for faster inference (3-4× speedup) vs FP16. Integrates with transformers and PEFT for QLoRA fine-tuning.
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.