DaFum

Use when asked to create, improve, review, migrate, sync, or validate repository context files for coding agents, including AGENTS.md, CLAUDE.md, CODEX.md, GEMINI.md, .cursorrules, or .github/copilot-instructions.md. Trigger on close synonyms and paraphrases such as "agent instructions", "coding agent setup", "AI-friendly repo", or "Copilot/Codex/Claude instructions".

Use when asked to create, improve, review, migrate, sync, or validate repository context files for coding agents, including AGENTS.md, CLAUDE.md, CODEX.md, GEMINI.md, .cursorrules, or .github/copilot-instructions.md. Trigger on close synonyms and paraphrases such as "agent instructions", "coding agent setup", "AI-friendly repo", or "Copilot/Codex/Claude instructions".

Use when writing, revising, or reviewing TSDoc comments for TypeScript APIs, IntelliSense hover text, exported functions, classes, interfaces, type aliases, generic helpers, parameter docs, examples, deprecation notices, defaults, thrown errors, release tags, TypeDoc or API Extractor output, or documentation lint feedback.

Audit GitHub Actions workflow efficiency and recommend fixes to reduce CI minutes and costs.

Use when Neurotoxic test commands, CI test jobs, or runner-owned suites need diagnosis or optimization: slow/flaky/hanging/OOMing runs, runtime regressions, worker tuning, sharding, fixture duplication, suite topology, `pnpm run test`, `pnpm run test:all`, node:test, Vitest, Playwright, perf, or locale coverage.

Audit GitHub Actions workflow efficiency and recommend fixes to reduce CI minutes and costs.

Use when Neurotoxic test commands, CI test jobs, or runner-owned suites need diagnosis or optimization: slow/flaky/hanging/OOMing runs, runtime regressions, worker tuning, sharding, fixture duplication, suite topology, `pnpm run test`, `pnpm run test:all`, node:test, Vitest, Playwright, perf, or locale coverage.

Take Playwright screenshots of any Neurotoxic game scene (INTRO, MENU, OVERWORLD, PREGIG, GIG, POSTGIG, GAMEOVER, CLINIC), UI overlays, PixiJS canvas, modal states, and visual regression baselines. Trigger when capturing screenshots, documenting UI, recording before/after diffs, debugging visual glitches, or producing visual regression tests.

You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.

Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies

Use when you have a written implementation plan to execute in a separate session with review checkpoints

Use when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation

Use when executing implementation plans with independent tasks in the current session

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes

Use when starting feature work that needs isolation from current workspace or before executing implementation plans - creates isolated git worktrees with smart directory selection and safety verification

Use when starting any conversation - establishes how to find and use skills, requiring Skill tool invocation before ANY response including clarifying questions

Write high-quality AGENTS.md, CLAUDE.md, or CODEX.md context files for code repositories. Use this skill whenever the user asks to create, write, generate, improve, or review a repository context file for coding agents — including any mention of "AGENTS.md", "CLAUDE.md", "CODEX.md", "GEMINI.md", ".cursorrules", "copilot-instructions.md", "context file for agents", "agent instructions", "coding agent setup", or requests like "help my AI agent understand this repo". Also trigger when users say things like "initialize agent config", "set up Claude Code for my project", "write instructions for Codex/Copilot", "make my repo AI-friendly", "migrate my .cursorrules", "sync my AI config across tools", or "validate my AGENTS.md". This skill is informed by peer-reviewed research on what actually helps coding agents solve tasks, and covers the full multi-tool ecosystem including Claude Code, Codex, Copilot, Cursor, Gemini CLI, Jules, and Windsurf.

Implement bug fixes, features, and targeted improvements for the Neurotoxic codebase — a 3D narrative adventure game built with React, Three.js (R3F), Zustand, and Rapier physics. Trigger when: the user wants to fix bugs, add features, refactor code, or improve anything touching dialogue, inventory, quests, scenes, audio, physics, or state logic in this game codebase. Also trigger for performance fixes, test additions, scene wiring, new crafting recipes, or dialogue tree changes. Use this skill even when the request seems small — it encodes critical gotchas (silent failures, stale closures, persistence pitfalls) that prevent common mistakes. If the user mentions store.ts, a scene file, audio.ts, dialog_uebersicht.md, or any game mechanic, this skill applies.

Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.

Implement bug fixes, features, and targeted improvements for the Neurotoxic codebase. Trigger when: requested changes touch `dialogue`, `inventory`, `quests`, `scene` wiring, audio, physics, or store/state logic — or when small, repository-conforming features, performance tweaks, or refactors are desired. This skill encodes repository conventions (scene wiring, `src/store.ts` patterns, `dialog_uebersicht.md` dialogue trees, `audio.ts` ambient handling, and `three` + `@react-three/rapier` patterns), provides a safe implementation checklist, and emits ready-to-apply patches or draft PR summaries. Route to specialists when appropriate (audio, deep state-safety, UI styling, balancing).

**WORKFLOW SKILL** — Produce a reusable SKILL.md from a conversation, PR, or code change. Outputs a conservative draft, concise clarifying questions, and runnable example prompts tailored to this repository.

Implement bug fixes, features, and targeted improvements for the Neurotoxic codebase — a 3D narrative adventure game built with React, Three.js (R3F), Zustand, and Rapier physics. Trigger when: the user wants to fix bugs, add features, refactor code, or improve anything touching dialogue, inventory, quests, scenes, audio, physics, or state logic in this game codebase. Also trigger for performance fixes, test additions, scene wiring, new crafting recipes, or dialogue tree changes. Use this skill even when the request seems small — it encodes critical gotchas (silent failures, stale closures, persistence pitfalls) that prevent common mistakes. If the user mentions store.ts, a scene file, audio.ts, dialog_uebersicht.md, or any game mechanic, this skill applies.

Implement bug fixes, features, and targeted improvements for the Neurotoxic codebase — a 3D narrative adventure game built with React, Three.js (R3F), Zustand, and Rapier physics. Trigger when: the user wants to fix bugs, add features, refactor code, or improve anything touching dialogue, inventory, quests, scenes, audio, physics, or state logic in this game codebase. Also trigger for performance fixes, test additions, scene wiring, new crafting recipes, or dialogue tree changes. Use this skill even when the request seems small — it encodes critical gotchas (silent failures, stale closures, persistence pitfalls) that prevent common mistakes. If the user mentions store.ts, a scene file, audio.ts, dialog_uebersicht.md, or any game mechanic, this skill applies.

Write high-quality AGENTS.md, CLAUDE.md, or CODEX.md context files for code repositories. Use this skill whenever the user asks to create, write, generate, improve, or review a repository context file for coding agents — including any mention of "AGENTS.md", "CLAUDE.md", "CODEX.md", "GEMINI.md", ".cursorrules", "copilot-instructions.md", "context file for agents", "agent instructions", "coding agent setup", or requests like "help my AI agent understand this repo". Also trigger when users say things like "initialize agent config", "set up Claude Code for my project", "write instructions for Codex/Copilot", "make my repo AI-friendly", "migrate my .cursorrules", "sync my AI config across tools", or "validate my AGENTS.md". This skill is informed by peer-reviewed research on what actually helps coding agents solve tasks, and covers the full multi-tool ecosystem including Claude Code, Codex, Copilot, Cursor, Gemini CLI, Jules, and Windsurf.

Apply the Null Agent instruction model for literal, immediate execution using repository AGENTS.md semantics. Use when tasks require strict instruction-following behavior, scoped AGENTS.md adherence, minimal interpretation, direct tool-driven execution steps, explicit "being the null agent" behavior, full-role impersonation guidance, and a hardened transformation process.