Best practices and guidelines for building production-grade 2D games in Go using Ebitengine (v2). Activate when designing game architecture, state machines, cycle timing, WASM audio sync, camera viewports, shaders, or custom typography.
Use this skill when guiding the user through game concept design, defining core mechanics, conducting interactive game design interviews (/grill-me), or generating a Game Design Document (GDD.md) for 2D games.
Use this skill when designing, generating, or optimizing pure-code procedural 2D sprites, tilemaps, vector graphics, particle systems, or visual effects without external image files. Activate for any tasks involving procedural rasterization, color palette theory, 2D matrix transformations, order of operations (scale, rotate, translate), particle pools, sub-frame interpolation/easing curves, or bitmap pixel crafting in game development.
Use this skill when designing, implementing, or optimizing pure-code procedural audio synthesis engines, sound effects (SFX), or polyphonic background music (BGM) without external assets. Activate for any questions or tasks involving audio DSP math, FM synthesis, ADSR envelopes, sound chip emulation (YM2612, SPC700, PS1 SPU), chiptunes, multi-track audio mixing, or game sound system code.
Use this skill when developing, validating, slicing, or managing 2D sprite sheets, animation sequences, frame durations, tag loops, and Aseprite files (.ase/.aseprite) for Ebitengine games. Serves as the specialized Animator Agent role for game development teams.
Master orchestrator and request router for 2D game development in Go using Ebitengine and Gemini AI models. Activate when handling user requests for game creation, game architecture, asset generation (procedural code vs. generative AI audio/images), Go code quality, or WebAssembly deployment.
Generate high-fidelity 44.1 kHz stereo music and audio from text prompts or images using Google's Lyria 3 models (lyria-3-clip-preview, lyria-3-pro-preview) with ADC authentication.
Generate or edit images using Nano Banana models (gemini-3.1-flash-lite-image, gemini-3.1-flash-image, gemini-3-pro-image, gemini-2.5-flash-image) with ADC auth. Supports "banana" as a verb (e.g., "please banana this") and Anime Dani / Chibi Dani character consistency.
MANDATORY for professional software development, production codebases, and open source maintenance. Activate this skill whenever designing features, refactoring code, maintaining repository hygiene, planning releases, reviewing breaking changes, or resolving technical ambiguity. Enforces semantic versioning (with zero-debt 0.x evolution), broken window code hygiene (dead code/comment cleanup, no lint ignores, strict error handling), RFC/ADR decision pipelines, 2x2 task prioritization, grounded 7-tier evidence research, and pre-release quality gates. Activate even when the user asks simple engineering questions or requests codebase maintenance.
Use this skill when you need to find real-world code examples, learn how an unfamiliar third-party library, API, SDK, or framework is used, or translate code patterns from other languages. Trigger when implementing integrations, resolving SDK setup/usage issues, or bootstrapping features where the user asks "how do I use X?", "show me an example of Y", or "integrate Y into my project".
Activate when working with Go (golang) code. Includes tooling and best practices to generate code efficiently and increase code quality.
Use this skill to fetch, verify, and resolve the latest stable versions of software packages, dependencies, and Gemini models. Activate this skill whenever the user asks to add or upgrade dependencies, initialize a new project, edit dependency files (such as package.json, requirements.txt, go.mod, Cargo.toml, Gemfile), or configure Gemini models in application code. Do not guess version numbers or rely on outdated training weights; use this skill to check official registries for the current stable release. Emphatically activate this skill when encountering dependency resolution errors, package installation failures, or security audits requiring package upgrades.
Use this skill when writing, editing, refactoring, debugging, or testing Python code, or when configuring Python project environments. Activate on any request to add Python features, fix bugs, resolve import errors, run pytest suites, configure linter rules in pyproject.toml, or manage packages and dependencies using uv. Trigger this skill even if the user does not explicitly mention Ruff, uv, or formatting, but asks for Python-related modifications, script creation, or dependency updates.
A comprehensive developer guide for A2UI v0.9.1. Trigger this skill for ANY question, bug, or coding task related to Agent-Driven User Interfaces (A2UI), MCP integrations, or component authoring.
Analyze, author, and optimize Agent Skills according to the official Agent Skills specification. Activate when creating new skills, auditing existing skills, optimizing trigger descriptions, designing eval suites (evals.json), bundling self-contained scripts, configuring the agentskills MCP server (https://agentskills.io/mcp), or organizing progressive disclosure structures (scripts/, references/, assets/, evals/). Use whenever reviewing, refactoring, or grading skills.
Use this skill when executing complex, multi-component engineering tasks, large-scale refactorings, or full-stack features that can be decomposed into parallelizable work. MANDATORY: Trigger immediately upon any mention of the word "swarm" (case-insensitive) in relation to planning or executing a task. Trigger immediately for: full-stack development (frontend + backend), multi-service integrations, complex database migrations, API contract designs, or high-ambiguity technical challenges requiring dedicated research (SPIKEs). Do NOT use for simple, isolated single-file edits, minor bug fixes, or basic script modifications where spawning subagents would introduce unnecessary communication overhead.
Official style guide and compliance rules for the Google Developers Blog. Activate to draft, edit, review, and validate technical blog posts, articles, and reviews to ensure they meet legal standards, the correct professional tone, and readability targets.
Use this skill when authoring, modifying, or auditing interactive developer tutorials and self-paced guides in the Google Codelabs (.lab.md) format. Activate whenever the user requests building a "codelab", "tutorial", "hands-on guide", or "step-by-step developer lab", or mentions publishing with the Codelab Asset Tool (Claat). Trigger also when converting existing documentation, blog posts, or readme files into structured tutorials, or when fixing formatting and validation issues in existing .lab.md files.
Activate this skill when proposing significant architectural changes, documenting codebase refactorings, or resolving design/technical debates. Use this skill to author, update, and maintain lightweight Architecture Decision Records (ADRs) to preserve engineering context. This skill must be triggered when the system needs to capture the why behind high-impact, hard-to-change structural choices.
Activate this skill when proposing major features, exploring complex design alternatives under high ambiguity, or gathering technical consensus. Use this skill to draft, review, and progress collaborative Request for Comments (RFC) documents. Trigger this skill whenever a systematic design proposal is needed before drafting immutable Architecture Decision Records.