Skills
99,448 packaged capabilities and playbooks for agents. Install any of them with one command.
Development
Advanced type-safe RPC patterns for Electrobun desktop applications. Covers bidirectional main↔webview communication, type safety with TypeScript, error handling and validation, performance optimization, streaming data patterns, batch operations, retry strategies, event-based communication, shared type definitions, RPC middleware, request/response patterns, and complex data synchronization. Use this skill when implementing complex communication between main and webview processes, need type-safe RPC with full IntelliSense, handling large data transfers, implementing real-time updates, building type-safe APIs between processes, debugging RPC issues, optimizing RPC performance, or implementing advanced patterns like streaming, batching, or pub/sub. Triggers include "RPC", "main webview communication", "type-safe RPC", "bidirectional RPC", "RPC performance", "RPC error handling", "shared types", "process communication", or "IPC patterns".
Designs neural network architectures that respect validated symmetry groups, recommending architecture families (G-CNN, steerable CNN, e3nn), layer patterns, and implementation libraries. Use when you have validated symmetry groups and need equivariant architecture design, or when user mentions equivariant layers, G-CNN, e3nn, steerable networks, or building symmetry into a model.
Safely diagnose and fix ESLint errors, warnings, and configuration compatibility issues. Use when the user asks to run lint, explain ESLint errors, perform scoped auto-fixes, or migrate ESLint configuration; prefer the project's locked package manager and versions, pre-check before modifying, and do not automatically download the latest version or bulk-rewrite unrelated source code.
Lints, tests, simulates, and formats Fastly VCL code using the falco tool. Also serves as the authoritative VCL reference via the falco Go source, which implements Fastly's full VCL dialect. Use when validating VCL syntax, running VCL linting, testing VCL locally, simulating VCL request handling, formatting VCL files, writing VCL unit tests with assertions, debugging VCL logic errors, looking up VCL function signatures or variable scopes, understanding VCL subroutine behavior, or running `falco lint`/`falco simulate`/`falco test`/`falco fmt`. Also applies when working with VCL syntax errors, type mismatches in VCL, choosing which VCL subroutine to use, or setting up a local VCL development and testing environment.
Verify that a developer-run feature behaved correctly by analyzing HTTP traffic captured by Fiddler Everywhere. Always use this skill when a developer asks whether their feature's HTTP calls completed correctly, wants to see what requests a feature made, needs to debug a failed API call, is checking traffic after running a feature, wants to confirm what each endpoint returned, or asks whether anything in the traffic looks wrong — even if they don't use the word "verify" or "Fiddler". Summarizes the capture by endpoint and flags likely issues such as failed calls, missing follow-up requests, retries, auth failures, timeouts, and suspicious status-code patterns. Requires Fiddler Everywhere to be running with its MCP server enabled.
Diagnose and fix Flutter text display issues including text overflow, orphaned words, line breaking, truncation, multi-style spans, and text scaling. Use when text clips, leaves dangling single words, overflows containers, or requires rich inline spans; route typography tokens and visual hierarchy to flutter-ui-design and screen-level layout constraints to flutter-responsive-layout.
Use when managing a Forgejo or Gitea server from the terminal: issues, pull requests, repositories, file contents, labels, milestones, releases, webhooks, user settings, or any /api/v1 endpoint through a safe generic API command. Do not use this skill for unrelated requests; route to the nearest named specialist.
Use when you need to derive a "page inventory → feature points → business processes → business rules" chain from frontend code with evidence, especially when the project has more than one routing entry point, complex dynamic routes/permissions/backend menus, and the team tends to speculate or miss evidence links.
Invoke when the user is preparing for game release, needs a launch checklist, store submission guidance, go/no-go decision framework, or day-one patch planning. Covers Steam, itch.io, Epic, and console submission pipelines. Triggers on: "launch", "release", "store submission", "go/no-go", "day-one patch", "launch readiness". Do NOT invoke for sprint planning (use game-sprint-plan) or code review (use game-code-review). Part of the AlterLab GameForge collection.
Use when a Git merge or rebase produces code conflicts, especially when AI-automated conflict resolution is error-prone (picking the wrong side, losing a refactor, reverting to an older version) and semantic analysis plus logic verification is needed. Also applies when a multi-round rebase stops repeatedly and conflict files need to be aggregated across rounds. Triggers — 「解冲突」「处理冲突」「git-conflict-resolve」「解决 merge 冲突」「解决 rebase 冲突」「冲突解决」 / conflict resolve, resolve merge conflict, resolve rebase conflict.
Go concurrency patterns for high-throughput web applications including worker pools, rate limiting, race detection, and safe shared state management. Use when implementing background task processing, rate limiters, or concurrent request handling.
Grilling session that challenges your plan against the existing domain model, sharpens terminology, and updates documentation (CONTEXT.md, ADRs) inline as decisions crystallise. Use when user wants to stress-test a plan against their project's language and documented decisions.
Find deepening opportunities in a codebase, informed by the domain language in CONTEXT.md and the decisions in docs/adr/. Use when the user wants to improve architecture, find refactoring opportunities, consolidate tightly-coupled modules, or make a codebase more testable and AI-navigable.
Use this skill to install, uninstall, list, inspect, and reset Android apps via `adb install` (with `-r` reinstall, `-d` allow-downgrade, `-t` allow test packages, `-g` grant all runtime permissions, `--user` per-user install, `adb install-multiple` for split APKs) and the on-device `pm` tool (`pm list packages [-f|-d|-e|-s|-3|-i|-u]`, `pm path`, `pm clear`, `pm grant` / `pm revoke`, `pm enable` / `pm disable`, `pm uninstall [-k]`, `cmd package dump-profiles`). Calls out the canonical hermetic-reset pattern (`pm clear` wipes data; `am force-stop` does NOT) and the common install errors (`INSTALL_FAILED_USER_RESTRICTED`, `INSTALL_FAILED_VERSION_DOWNGRADE`, signing-conflict). Use when the user mentions "reset app between tests", `pm clear`, `am force-stop` confusion, install fails after debugger, split APK install, runtime permission grants, multi-user profile installs, or "how do I list third-party apps".
Guides iOS app development with Swift including SwiftUI views, UIKit view controllers, navigation, async/await networking, Core Data persistence, Xcode project configuration, and App Store submission. Use when the user asks about iOS Swift development, needs to create iOS applications, implement SwiftUI or UIKit components, or configure Xcode signing and TestFlight.
Use when a coding agent works in a repository where Jujutsu (jj) is available (a `.jj/` directory) — running jj commands, making speculative edits, splitting changes, recovering with the operation log (jj undo / jj op restore), handing off a PR to Git/GitHub/GitLab, updating a PR after review, or coordinating parallel agents. Uses jj as the local change-management layer while keeping Git as the canonical remote, PR, CI, and audit interface. Also use to avoid jj's agent footguns (pager/editor hangs, commit absorption, colocated git/jj desync) or to replace fragile git reset/checkout/stash/rebase recovery with reversible jj operations.
Guide for llama.cpp, the C/C++ LLM inference framework by ggml-org. Covers the C API (llama.h), GGUF format, quantization (Q4_K_M, Q8_0, IQ4_XS), CMake builds, GPU backends (CUDA, Vulkan, Metal, ROCm), HTTP server with OpenAI-compatible API, embeddings, grammar constraints, function calling, LoRA, speculative decoding, multimodal, and UE5 integration. Use when: llama.cpp, GGUF models, local LLM inference, llama.h, llama-server, quantizing, ggml, building/linking llama.cpp, GPU acceleration, llama.cpp embeddings, grammar/JSON output, llama.cpp in Unreal Engine, llama_* API functions, GGUF format, converting HuggingFace to GGUF, or comparing with vLLM/Ollama/TensorRT-LLM.
Use when writing MATLAB functions with arguments blocks — repeating arguments (arguments (Repeating)), .?ClassName property import in constructors, name-value forwarding with namedargs2cell, or migrating from inputParser or validateattributes. Also when reviewing signatures for implicit-conversion pitfalls (size reshaping, class coercion, computed defaults), or when restricting, constraining, or validating function inputs — scalar vs vector enforcement, type rejection, size checking. Also when asked to harden inputs, make a function more robust, tighten input checking, or rewrite for safer input acceptance — even without "validation" or "arguments block" wording. Triggers on: arguments block, mustBe validators, name-value arguments, varargin, inputParser, nargin/narginchk, restrict input, type checking, scalar constraint, harden inputs, defensive programming. Not for: App Designer callbacks, Simulink mask parameters, class inheritance, or runtime validation outside function or property declarations.
Use when you need a runnable PyTorch training repo for a medical-imaging task (segmentation, classification, detection, synthesis, self-supervised, or fine-tuning a pretrained backbone). Emits a patient-level seed-locked split, train/evaluate scripts, config and a Methods stub.
Write Python code in n8n Code nodes. Use when writing Python in n8n, using _input/_json/_node syntax, working with standard library, or need to understand Python limitations in n8n Code nodes. Use this skill when the user specifically requests Python for an n8n Code node. Note — JavaScript is recommended for 95% of use cases — only use Python when the user explicitly prefers it or the task requires Python-specific standard library capabilities (regex, hashlib, statistics). EXCEPTION — for Python in the AI-agent-callable Custom Code Tool (@n8n/n8n-nodes-langchain.toolCode), use the n8n-code-tool skill instead (input is _query, return must be a string).
Validate n8n expression syntax and fix common errors. Use when writing n8n expressions, using {{}} syntax, accessing $json/$node variables, troubleshooting expression errors, mapping data between nodes, or referencing webhook data in workflows. Use this skill whenever configuring node fields that reference data from previous nodes — expressions are how n8n passes data between nodes, and getting the syntax wrong is the most common source of workflow errors. Also use when asked whether a complex expression hurts performance.
Interact with Obsidian vaults using the Obsidian CLI to read, create, search, and manage notes, tasks, properties, and more. Also supports plugin and theme development with commands to reload plugins, run JavaScript, capture errors, take screenshots, and inspect the DOM. Use when the user asks to interact with their Obsidian vault, manage notes, search vault content, perform vault operations from the command line, or develop and debug Obsidian plugins and themes.
Integrate OnboardJS into React projects for building user onboarding flows. Use when: (1) Setting up OnboardJS in a React/Next.js project (2) Creating multi-step onboarding, wizards, or guided flows (3) Implementing conditional navigation, data persistence, or step validation (4) Working with OnboardingProvider, useOnboarding hook, or step components Triggers: "onboarding", "onboardjs", "wizard flow", "user onboarding", "guided tour", "multi-step form"
Operate Orca-managed worktrees, folder contexts, terminals, repos, automations, artifacts, skill sharing, worktree comments, and Orca's embedded browser through the `orca` CLI. Use when the user says "$orca-cli", "Orca worktree", "child worktree", "spawn codex/claude in a worktree", "read/wait/send Orca terminal", "handoff" / "handover" / "give this to another agent", "Orca browser", "orca artifacts", or "share skills". Prefer it over raw git worktree, ad hoc PTYs, or Computer Use when Orca state is involved. Use Computer Use only when a visible window needs GUI control that a CLI, filesystem, or API cannot do.
Version-pinned inventory of the telemetry (spans, metrics, logs, attributes) emitted by OpenTelemetry collector components and SDK instrumentation packages. Use when working with a covered component — what it emits at a given version, or how emission changed across versions — and when upgrading a component or SDK version, to see the telemetry after the change.
Drive a live browser from the CLI with playwright-cli to navigate, interact, snapshot, and capture evidence. Use for ad-hoc browser commands, page inspection, screenshots, traces, network mocking, session management, or interactive debugging—not authoring @playwright/test specs. Keywords: playwright-cli, browser automation, live session, snapshot, console, tracing.
Engineer effective LLM prompts using zero-shot, few-shot, chain-of-thought, and structured output techniques. Use when building LLM applications requiring reliable outputs, implementing RAG systems, creating AI agents, or optimizing prompt quality and cost. Covers OpenAI, Anthropic, and open-source models with multi-language examples (Python/TypeScript).
Unvarnished technical criticism combining Linus Torvalds' precision, Gordon Ramsay's standards, and James Bach's BS-detection. Use when code/tests need harsh reality checks, certification schemes smell fishy, or technical decisions lack rigor. No sugar-coating, just surgical truth about what's broken and why.
Execute small, well-understood changes directly in the current session with minimal ceremony and ponytail taste. Use when explicitly requested, or automatically for localized fixes, mechanical edits, config tweaks, and short scripts with clear behavior and a known verification path. No plan documents or grilling; do not add tests by default. Prefer steady-do when design, compatibility, or acceptance decisions remain; prefer plan-brief when planning and execution need a separate-session handoff.
Reverse engineering binaries using radare2 (r2). Analyze executables, disassemble code, find functions, extract strings, patch binaries, debug programs, and explore firmware. Supports persistent sessions for large binaries to avoid re-analysis. Use when the user wants to reverse engineer, disassemble, analyze, debug, or patch any binary file including ELF, PE, Mach-O, firmware, or shellcode.
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.