Claude Skills by gaelic-ghost

Plan and validate .NET SDK, target framework, package, and language-version upgrades for F#, C#, and mixed solutions with compatibility checks and staged validation.

Choose the right Apple platform game-development stack before implementation. Use when Codex needs to route a game task across SpriteKit, SceneKit, RealityKit, Metal, GameplayKit, Game Controller, Core Haptics, GameKit, SwiftUI/AppKit/UIKit host surfaces, assets, profiling, or Apple Dev Skills handoffs.

Guide Core Haptics and game feedback work for Apple games. Use when Codex designs, implements, repairs, or validates CHHapticEngine, CHHapticPattern, audio-haptic feedback, controller haptics, capability checks, fallback behavior, accessibility-sensitive feedback, or physical-device haptic validation.

Guide Apple Game Controller framework input work for games. Use when Codex works on GCController, GCDevice, extended gamepad mappings, virtual controllers, keyboard and mouse game input, controller connection lifecycle, controller haptics handoffs, input evidence, accessibility alternatives, or device-aware validation.

Select and operate Apple Game Porting Toolkit 3 or 4 for bringing an existing Windows game or engine to Apple platforms. Use when Codex encounters a Windows executable, Direct3D, Vulkan, DXIL, Metal Shader Converter, GPTK evaluation, a porting milestone, or the upstream game-porting-skills plugin; route native Metal renderer work to metal-game-rendering-workflow.

Guide GameplayKit simulation and gameplay-architecture work. Use when Codex works on GKEntity, GKComponent, GKComponentSystem, GKState, GKStateMachine, GKGraph, GKAgent, GKGoal, GKBehavior, GKRandom, SpriteKit or SceneKit integration, pathfinding, agents, state machines, reusable gameplay logic, or game simulation validation.

Guide Metal game asset residency and streaming for textures, meshes, mipmaps, heaps, residency sets, sparse resources, and fast resource loading. Use when Codex investigates GPU memory pressure, streaming hitches, resource lifetime, texture or mesh upload, sparse texture mapping, or asset budgets in a Metal renderer; do not use for generic asset-catalog or SceneKit resource work.

Guide native Apple game and custom-renderer work with Metal 3 and Metal 4. Use when Codex owns MTKView, CAMetalLayer, drawable presentation, render passes, command encoding, MSL shaders, pipeline states, GPU resources, Metal capability fallback, GPU capture, or renderer validation; do not use for DirectX or Vulkan porting, which belongs to the Game Porting Toolkit workflow.

Plan and implement experimental Metal neural-rendering paths with MetalFX, Metal 4 machine-learning passes, Metal tensors, and Metal Performance Primitives. Use when Codex works on learned tone mapping, neural denoising, neural materials, inline tensor operations, MTL4MachineLearningCommandEncoder, or model inference scheduled beside rendering; preserve a conventional rendering fallback.

Guide native Apple-game MetalFX integration for spatial or temporal upscaling, frame interpolation, and temporal denoising. Use when Codex works with MetalFX descriptors, jitter, motion vectors, depth, exposure, history, dynamic resolution, present threads, or MetalFX performance validation; use Game Porting Toolkit guidance for a D3D or Vulkan port.

Guide SceneKit game implementation, repair, and maintenance. Use when Codex works on SCNScene, SCNView, SCNNode, cameras, lights, materials, geometry, animation, 3D physics, .scn files, .scnassets, model assets, SceneKit and SwiftUI/AppKit/UIKit integration, or RealityKit migration handoffs.

Guide SpriteKit game implementation, repair, and maintenance. Use when Codex works on SKScene, SKView, SKNode, actions, particles, cameras, constraints, 2D physics, tile maps, SpriteKit resources, .sks files, SpriteKit and GameplayKit integration, or SpriteKit validation in Apple game projects.

Guide Apple game profiling and performance-evidence work in Xcode and Instruments. Use when Codex investigates frame pacing, stutter, FPS, CPU/GPU overlap, Time Profiler evidence, Game Performance or Game Memory templates, Metal Performance HUD routing, memory pressure, thermal state, trace capture, or profiling handoffs for SpriteKit, SceneKit, RealityKit, or Metal-backed games.

Route an Apple communication request to iMessage collaboration, communication notifications, Push to Talk, VoIP calling, default-app, or macOS companion workflows before implementation.

Choose a supported chat, calling, business-messaging, or collaboration integration before implementation. Use for Discord, Telegram, Slack, Teams, WhatsApp Business, SMS/MMS/RCS, Google Meet, iMessage collaboration, and Apple VoIP.

Implement Apple communication and Time Sensitive notifications with intent donation, Focus behavior, and a bounded Notification Service Extension processing path.

Plan conversation identity, state, consent, escalation, transcript, and human-handoff behavior for bots and business messaging agents.

Plan iOS and iPadOS default messaging, calling, dialer, and carrier messaging app roles while separating documented system integrations from an app-owned macOS companion implementation.

Plan, build, and validate Discord apps, bots, commands, interactions, Gateway event handling, and Activities using current Discord documentation.

Plan and validate Google Meet add-ons, conference REST integrations, event subscriptions, artifacts, and preview media API work.

Build an iMessage app or collaboration feature with Messages extensions, interactive message sessions, and Shared with You without implying server-side access to personal conversations.

Build an Apple Push to Talk channel experience with system controls, ephemeral PTT pushes, audio-session ownership, and an app-owned communication backend.

Plan, build, and validate Slack workspace apps with OAuth, Events API, interactivity, slash commands, modals, and Socket Mode.

Plan, build, and validate consent-aware SMS, MMS, and RCS business messaging agents, including sender setup, webhooks, rich content, delivery, and fallback.

Plan, build, and validate Microsoft Teams bots and agents with Teams SDK or Microsoft 365 Agents SDK, Microsoft Entra, Adaptive Cards, and tenant deployment boundaries.

Plan, build, and validate Telegram Bot API integrations, webhooks or polling, inline interactions, and Mini Apps.

Build a native Apple VoIP calling app by separating CallKit or LiveCommunicationKit, PushKit, AVFAudio, SIP or WebRTC transport, and call-state ownership.

Design and repair secure inbound events, webhooks, interactions, retries, acknowledgements, and delivery observability for messaging and collaboration platforms.

Plan and validate official WhatsApp Business messaging integrations, including onboarding, webhooks, conversation policy, templates, consent, and provider boundaries.

Reproduce and evaluate refusal-direction ablation on controlled checkpoints. Use when estimating refusal representations, applying activation or weight orthogonalization, measuring behavior change, or comparing ablated artifacts.

Benchmark model runtimes across latency, throughput, memory, energy, load time, size, and stability. Use when comparing Core AI, Core ML, MLX, ExecuTorch, PyTorch, quantization, devices, or packaging.

Compare and select Core AI, Core ML, MLX, MLX Swift, MLX LM, ExecuTorch Apple delegates, or Foundation Models. Use when an Apple model workflow needs a runtime choice and implementation handoff.

Route language-model training, data, evaluation, checkpoint, representation, steering, ablation, jailbreak, tool-calling, Apple-runtime, and benchmark requests. Use when the primary workflow or Socket owner is unclear.

Compare base, adapter, merged, quantized, converted, or intervention-derived checkpoints. Use when selecting an artifact, investigating regression, verifying packaging, or balancing quality, safety, size, and runtime.

Design a reproducible model experiment with hypotheses, controls, baselines, metrics, budgets, provenance, artifacts, and stop conditions. Use before model work whose results must support a decision.

Evaluate authorized jailbreak and prompt-injection resilience with reproducible attacks, budgets, graders, and side-effect boundaries. Use for controlled checkpoints, local harnesses, or explicitly authorized systems.

Design repeatable language-model capability, quality, behavior, regression, or safety evaluations. Use when creating cases and graders, comparing a treatment with a baseline, or deciding whether a checkpoint passes.

Evaluate tool selection, arguments, schemas, execution, observation use, recovery, and side effects. Use when comparing tool-capable checkpoints, prompts, parsers, agent loops, function schemas, or harnesses.

Plan supervised fine-tuning, LoRA, QLoRA, or full-parameter training with reproducible evaluation gates. Use when adapting a checkpoint or choosing precision, adapters, optimization, checkpointing, or resume behavior.

Prepare language-model training, preference, red-team, or evaluation datasets. Use when collecting, filtering, deduplicating, templating, splitting, or auditing examples for provenance, leakage, consent, or sensitive data.

Design causal research into model activations, features, attention, residual streams, probes, and circuits. Use when locating behavior, testing a direction or feature, comparing layers, or reproducing interpretability research.

Implement and evaluate activation steering, representation engineering, logit changes, or weight-space interventions. Use when changing model behavior without ordinary fine-tuning or sweeping layer, strength, and persistence choices.

Choose the right application transport or real-time protocol before implementation. Use when Codex needs to route work across plain HTTP routes, HTTP/2, HTTP/3, QUIC, WebSocket, WebTransport-adjacent designs, WebRTC media or data channels, Media over QUIC, gRPC, RPC, streaming APIs, mobile app networking, browser networking, server frameworks, deployment proxies, or protocol diagnostics.

Plan, implement, test, and diagnose HTTP/3 or QUIC work, including RFC-backed source checks, ALPN, Alt-Svc, TLS, UDP reachability, QUIC streams and datagrams, HTTP semantics over QUIC, proxy/CDN/server support, qlog or packet evidence, and handoffs to stack-specific implementation plugins.

Diagnose modern network protocol failures with evidence, including HTTP version negotiation, QUIC and UDP reachability, TLS and ALPN, Alt-Svc, browser networking, WebRTC ICE/STUN/TURN, MoQ relay logs, proxies, CDNs, firewalls, NAT, packet captures, qlog, server logs, and stack-specific handoffs.

Plan, evaluate, and diagnose Media over QUIC or MoQ work, including draft-state checks, relay/client/server topology, publish/subscribe object flow, media packaging, live latency, reliability tradeoffs, authorization/privacy drafts, qlog or protocol evidence, and fallback routing to WebRTC, HLS/DASH, HTTP streaming, or stack-specific implementation plugins.

Plan, implement, test, and diagnose WebRTC work, including W3C and IETF source checks, browser and native runtime support, signaling boundaries, ICE/STUN/TURN, DTLS, SRTP, SDP offer/answer, tracks, transceivers, data channels, media devices, permissions, NAT traversal, SFU/MCU decisions, and handoffs to web, mobile, server, or deployment plugins.

Search and triage technology jobs through Dice's remote MCP server. Use when the user wants Dice.com job search help, MCP setup guidance, role-fit triage, resume-to-listing comparison, saved search prompt design, salary or skills trend checks, or recurring job-search automation boundaries grounded in Dice's documented search_jobs tool.

Diagnose Python uv sync, lock, import, test, Ruff, mypy, FastAPI, FastMCP, packaging, and CI failures with concrete phase classification and next checks.

Design and maintain Python CI workflows around uv, pytest, Ruff, mypy, package build checks, dependency caching, Python version matrices, and local-command parity.