Claude Skills by vasilyu1983

Configures Claude Code hooks and Codex hooks.json/notify callbacks. Use when adding guardrails, preflight, audit trails, worktree automation, or budget enforcement.

Configures and hardens MCP servers for Claude Code and Codex agents. Use when connecting databases, APIs, files, or SaaS via MCP, or building custom servers.

Manages AGENTS.md, CLAUDE.md, and scoped repo rules for Claude Code and Codex. Use when fixing stale memory, ignored instructions, memory audits, or model-upgrade migration.

Adds per-skill learnings loops for dated patterns, mistakes, and domain facts. Use when wiring skill memory, consolidation, or drift audits.

Creates and audits agent skills with SKILL.md, references, scripts, and platform-scoped metadata. Use when creating, updating, or validating shared skills.

Coordinates multi-agent execution across subagents, teams, and workflows. Use when planning dependency-aware fan-out, verifier passes, runtime selection, or Loop Engineering.

AI agent architecture, graph and loop composition, protocol choice, evaluation, and observability. Use when scoping or reviewing systems before implementation.

Chooses among AI/ML approaches: classical ML, LLM, RAG, fine-tuning, agents, multimodal, embeddings/recsys, dense/MoE/SSM/diffusion. Use when picking or scaling an architecture.

Designs slash-command runtimes for coding-agent CLIs. Use when modeling command registries, lazy loading, aliases, forked commands, or remote-safe dispatch.

Designs execution sandboxes for coding agents. Use when modeling process isolation, filesystem policy, network controls, workspace mounts, or destructive-command boundaries.

Designs coding-agent observability and evals. Use when measuring traces, replay, checkpoint lineage, quality trajectories, tool grading, regression, or cost.

Designs approval and permission systems for coding-agent runtimes. Use when modeling tool approvals, plan-mode transitions, sandbox prompts, or worker permission handoffs.

Designs plugin systems for coding-agent runtimes and CLIs. Use when adding plugin manifests, extension points, built-in plugins, or reloadable agent integrations.

Designs provider runtimes for coding agents. Use when modeling model abstraction, streaming semantics, tool-call normalization, retries, or fallback routing.

Designs release and distribution systems for coding-agent CLIs. Use when modeling packaging, auto-update channels, plugin compatibility, cache migrations, or install footprints.

Designs remote execution and bridge runtimes for coding agents. Use when implementing remote sessions, local-UI remote-execution, reconnect logic, or permission bridging.

Designs session lifecycle for coding-agent runtimes. Use when implementing resume, transcript restoration, checkpoint rewind, cross-worktree recovery, or session-state persistence.

Designs settings and policy layers for coding-agent runtimes. Use when modeling source precedence, managed policy, env controls, or runtime settings validation.

Designs task runtimes for Loop Engineering, Graph Engineering, and background work. Use when work needs task lists, cyclic/workflow graphs, cancellation, or teammate coordination.

Designs terminal-first coding-agent UX: REPL, prompt input, status lines, keybindings, display modes. Use when shaping TUI rendering, history, or background-task navigation.

Designs tool runtimes for coding agents. Use when modeling tool registries, deferred loading, permission-aware execution, tool search, or remote tool rendering.

Creates coding agents on Claude Code, Codex, and Agent SDK. Use when defining review, test, refactor, or team agents — not building a runtime.

Builds and audits LLM pretraining corpora: extraction, filtering, dedup, decontamination, data mixing, synthetic data. Use when curating or ablating a pretraining data pipeline.

Builds repeatable deep-research workflows for verified synthesis. Use when producing evidence-backed briefs, comparisons, dossiers, or research pipelines.

Guides multi-GPU pre-training: DDP, FSDP2, ZeRO, tensor/pipeline/expert parallelism, fp8/Muon. Use when scaling a run, training MoE, or reproducing GPT-2 on rented GPUs.

Designs trustworthy LLM, agent, responsible-AI, and multimodal evaluations. Use when measuring quality, fairness, privacy, grounding, safety, or judge reliability.

LLM inference patterns for latency, batching, caching, quantization, routing, and serving stacks. Use when optimizing throughput, tail latency, or serving cost.

Guides the LLM lifecycle from strategy to deployment. Use when planning, comparing, fine-tuning, distilling, compressing, migrating, or operating LLM systems.

Runs local and self-hosted LLM workflows with Ollama, LM Studio, MLX, Open WebUI, llamafile, and adapters. Use when operating private model stacks.

Builds ML, responsible-AI, and multimodal models. Use when doing data science or explaining fairness, privacy, speech, vision-language, or diffusion mechanics.

Time-series forecasting with temporal validation, panel models, probabilistic forecasts, and TS foundation models. Use when modeling ordered observations.

Operates ML, LLM, responsible-AI, and multimodal systems. Use when deploying, monitoring, governing, or responding to production AI failures.

Post-training and alignment: reward modeling, RLHF/PPO, DPO/DAAs, GRPO, RLVR, RLAIF, over-optimization. Use when adapting an SFT model with preference or verifiable-reward signals.

Builds a transformer/GPT and BPE tokenizer from scratch. Use when implementing autograd, self-attention, a nanoGPT-style pretraining loop, or a byte-level tokenizer.

Designs operating models for AI in product teams. Use when planning platform ownership, provider strategy, data boundaries, evals, or sensitive-data controls.

Prompt engineering for production LLMs — structured outputs, evals, RAG, tool workflows, multimodal prompting, and safety. Use when designing, debugging, or shipping prompts.

Designs retrieval-augmented generation and search systems. Use when choosing retrieval, chunking, hybrid search, grounding, or RAG evaluation patterns.

Sizes models and token budgets using Kaplan/Chinchilla scaling laws. Use when reasoning about compute-optimal N and D, tokens-per-parameter ratios, or over-training tradeoffs.

Builds vector-brain implementations for repos, docs hubs, and compliance corpora. Use when creating pgvector retrieval brains with scripts, SQL, manifests, and evals.

Builds production voice bots and IVR with Python STT/TTS pipelines. Use when designing telephony, streaming audio, latency budgets, or voice quality monitoring.

Builds analytics engineering layers for metrics, contracts, and BI-ready models. Use when shaping dbt or SQLMesh marts, metric governance, lineage, or data quality.

Designs lakehouse platforms across Iceberg, Delta, Hudi, and Paimon. Use when choosing catalogs, CDC paths, query engines, governance, or cost controls.

Automates Metabase cards, dashboards, Remote Sync, embedding, tenants, and the Agent API/MCP server for AI workflows. Use when scripting, promoting, or embedding Metabase content.

Diagnoses and tunes SQL for OLTP workloads on PostgreSQL, MySQL, and SQL Server. Use when tuning queries, reading plans, indexing, or fixing lock contention.

Designs streaming platforms for Kafka, Flink, CDC, and lakehouse ingestion. Use when planning event backbones, CDC pipelines, schema governance, or real-time lakehouse delivery.

Measures AI coding impact and extension robustness. Use when tracking delivery, quality trajectories, cost, experience, pilots, scorecards, or leadership reporting.

Designs durable API contracts across REST, GraphQL, gRPC, tRPC, and AsyncAPI. Use when specifying interfaces, auth, versioning, errors, rate limits, or agent APIs.

Builds per-repo code graphs in JSON and markdown-ready derived artifacts. Use when you need blast radius, symbol-level maps, import graphs, inheritance, or test links.

Context-driven AI development with AGENTS.md, repo knowledge bases, Claude Code, Codex, and Copilot. Use when adopting repo-native AI workflows or multi-repo setups.

Builds multi-repo context hubs and compiled markdown knowledge maps. Use when profiling repo portfolios or assembling LLM-ready cross-repo knowledge bases.