dotfiles

Use when implementing real-time DSP algorithms, audio thread architecture, or offline spectral analysis in C++/Rust. Covers filters, STFT pipelines, lock-free concurrency, and RT safety. Not for JUCE plugin lifecycle wiring or ffmpeg.

Use for real-time audio code safety, determinism, and numeric hygiene. Required foundation for DSP, audio analysis, audio systems, and JUCE work. Not for game-audio middleware or ffmpeg/video tasks.

Use when programmatically processing video/audio with libffmpeg C API. Not for command-line ffmpeg operations.

Use for JUCE audio apps/plugins: AudioProcessor lifecycle, VST3/AU targets, parameter systems, and thread separation. Requires audio-engineering-principles. Not for game-audio middleware or non-JUCE frameworks.

Use when writing, reviewing, or refactoring C++ code — covers workflow, essential patterns, and pre-commit checklist for modern C++17/20.

Use when writing or fixing C++ tests, configuring GoogleTest/CTest, or adding coverage/sanitizers.

Use for .NET 8+ C# apps: ASP.NET Core APIs, Blazor, EF Core, minimal APIs, async patterns, CQRS with MediatR, and Dapper.

Use when setting up or using shadcn/ui components (install, configure, themes, forms, dialogs, tables). Not for non-React UI frameworks.

Use for Tailwind v4 design systems: tokens, component libraries, and responsive patterns.

Use for Vite config, plugins, SSR, or Vite 8/Rolldown migration. Not for Vitest testing.

Use for Vitest tests: mocking, coverage, fixtures, filtering, and setup. Not for non-Vitest testing frameworks.

Use when making system design decisions in Godot 4.x — ownership, signal topology, performance trade-offs, scene boundaries, or large-scale feature architecture. Not for routine scripting tasks (use godot-core).

Use for Godot game-audio and middleware integration — AudioBus routing, Wwise/FMOD events, adaptive music, and procedural sound systems. Not for DSP, JUCE, ffmpeg, or offline audio pipelines.

Use when writing GDScript, creating scenes, adding nodes, or doing day-to-day Godot 4.x development. Covers project structure, script patterns, signals, resources, and MCP tools.

Use for security audits, vulnerability remediation, hardening configs, and secure coding patterns. Not for compliance frameworks or policy writing.

Use for Dockerfiles, multi-stage builds, image optimization, Compose, container security, and runtime debugging. Not for Kubernetes/Helm or cloud-specific container services.

Use when deploying or managing Kubernetes workloads requiring cluster configuration, security hardening, or troubleshooting. Invoke for Helm charts, RBAC policies, NetworkPolicies, storage configuration, performance optimization.

Use when writing, reviewing, or refactoring Terraform HCL — covers project setup, resource authoring, module development, and pre-commit validation.

Use for Rails apps: ActiveRecord/models, migrations, controllers/routes, background jobs, caching, and Hotwire. Not for non-Rails Ruby.

Use for async Rust with Tokio: tasks, channels, error handling, concurrency patterns, and debugging async issues.

Use when writing or refactoring Rust for idioms, ownership/borrowing choices, Result-based error handling, performance, and tests/docs.

Use when working with PostgreSQL — schema design, indexing, query optimization (EXPLAIN/ANALYZE), connection pooling, VACUUM/autovacuum, migrations, and performance troubleshooting.

Use when working with SQLite databases — schema design, migrations, query tuning, WAL/locking, pragma configuration, and file-based deployment.

Use for automated browser tasks (navigation, form filling, scraping, screenshots, web app testing). Not for simple one-off browsing.

Use when asked to write or review AGENTS.md/CLAUDE.md/LLMs.md context files.

Use before creative work — features, components, or behavior changes. Explores intent before implementation.

Use to map or understand a repository — codemaps, architecture overviews, or codebase documentation. Triggers: "map the codebase", "codemap", "architecture overview".

Use when conducting code reviews or establishing review standards. Not for requesting or receiving review.

Use after AI-generated code to remove slop (excess comments/defensive checks). Not for general refactors.

Use when implementation is complete and you need to decide how to integrate — merge, PR, or cleanup.

Use when receiving code review feedback before implementing — rigor required, not blind agreement.

Use after manual code changes to improve clarity/readability without changing behavior. Not for AI slop cleanup.

Use when starting feature work needing isolation, or before executing implementation plans.

Use when starting any conversation — establishes skill discovery and loading workflow.

Use before claiming work is complete — run verification commands and confirm output before assertions.