btgaskin

Guide for operating, extending, and interpreting BrainlessLab.jl, the Julia research platform for behaviour from self-organising neural substrates. Use for every task in the brainless-lab repository. Covers CompositionSpec, EvaluationSpec, typed registries, profile, sweep, ablation, evolution, benchmark plans, ExperimentSpec, versioned records, evidence boundaries, and node or task extension. Pair with the Julia skill for language, dispatch, inference, allocation, and package hygiene.

Comprehensive guide for writing, reviewing, debugging, and optimizing Julia code for scientific computing — covering type stability, multiple dispatch design, memory and allocations, profiling/debugging tools, package and environment management, parallelism, and the scientific ecosystem (SciML/DifferentialEquations.jl, Agents.jl, DynamicalSystems.jl) relevant to computational neuroscience and artificial-life modeling. Use this skill any time Julia code is being written, generated, reviewed, refactored, explained, or debugged — including diagnosing why Julia code is slow, fixing type instabilities, choosing a package or data structure, setting up a Julia project environment, or reviewing someone else's (or your own previously generated) Julia code for correctness and idiom — even if the request doesn't explicitly mention "performance," "Julia," or this skill by name.

Create a personalized study plan through structured interview, research, and adaptive scheduling. Use when starting a new learning project, preparing for exams or interviews, or building a structured self-study program. Triggers: study plan, learning plan, exam prep, interview prep, create study schedule, learn new topic, prepare for certification, technical interview prep.

Run a daily study session with spaced repetition, coding exercises, and progress tracking. Loads a study plan and adapts each session based on schedule, due cards, and progress. Use when ready to study, practice, review flashcards, or work through a study plan. Triggers: study session, daily study, practice, review cards, study time, let's study, drill, prep session.

Build production AI agents with the Claude Agent SDK. Use when working with: (1) Creating autonomous agents with query() or ClaudeSDKClient, (2) Defining custom tools via in-process MCP servers, (3) Implementing hooks for tool interception and permissions, (4) Building multi-agent systems with subagents, (5) Managing sessions and state across queries, (6) Integrating MCP servers for external capabilities. Covers Python and TypeScript SDKs with production patterns.

Write correct, idiomatic Apple MLX code for Apple Silicon ML. Use when working with MLX arrays, neural networks, training loops, lazy evaluation, unified memory, mx.eval, mx.compile, Metal GPU, memory optimization, quantization, or Apple Silicon performance. Covers critical API differences from PyTorch/NumPy, array indexing gotchas (lists must be mx.array, slices create copies), NHWC format for Conv2d, __call__ not forward(), float64 CPU-only, mlx-lm integration, and debugging patterns.

Use when planning or running an end-to-end literature review with this framework. Guides question framing, search-term design, PRISMA/PRISMA-S reporting, config drafting, pilot sampling, QA gates, rule versioning, PDF handling, and failure-mode safeguards.