Assesses whether a business question can be answered with data available in a Bauplan lakehouse. Maps business concepts to tables and columns, checks data quality on the relevant subset, validates semantic fit, and renders a verdict: answerable, partially answerable, or not answerable. Produces a structured feasibility report. Use when a user brings a business question, asks 'can we answer this', wants to know if the data supports an analysis, or before building a one-off analysis or pipeline.
Apply when scoping, reviewing, or documenting cross-cutting VTEX commerce architecture across storefront, IO, headless, marketplace, payments, or any other VTEX module. Grounds work in the Well-Architected Commerce framework—Technical Foundation (reliability, trust, integrity; security, infrastructure, compliance), Future-proof (innovation, simplicity, efficiency; scalable and adaptable solutions), and Operational Excellence (accuracy, accountability, data-driven improvement; process and customer experience). Routes implementation detail to product tracks (IO caching and paths, Master Data strategy, marketplace integrations). Use for solution design, architecture reviews, and RFP-level technical structure.
Guide for querying and operating on data through the Agentic Data Protocol (ADP) MCP tools. This skill teaches how to use adp_discover, adp_describe, adp_validate, and adp_execute tools effectively. Use this skill whenever the user wants to explore available data resources, query data, look up records, insert new data, or update existing data through ADP — even if they don't mention "ADP" explicitly. Triggers include "find data", "query records", "look up customer", "insert rows", "update entries", "what data is available", "search for similar items", "show me the schema", or any data exploration and manipulation task where ADP MCP tools are connected. --- # ADP Skill — Data Access via Agentic Data Protocol This skill helps you interact with data through four ADP MCP tools that connect to an ADP Hypervisor. The Hypervisor is a policy-enforcing gateway that abstracts away backend differences (SQL databases, MongoDB, vector stores, file systems) behind a single intent-based interface. ## Core Workflow Always follow this sequence — each step builds on the previous one: ``` discover → describe → validate (optional) → execute ``` 1. **Discover** — Find what data resources are available 2. **Describe** — Read the usage contract for a specific resource (field schema, allowed operators, required predicates) 3. **Validate** — Dry-run your intent to catch errors before execution (recommended but optional) 4. **Execute** — Run the intent and get results Skipping `describe` leads to malformed intents because you won't know which fields exist, which predicates are required, or which operators are allowed. Always describe before building an intent. ## The Four MCP Tools ### 1. `adp_discover` — Browse Available Resources Find what data is available. Use filters to narrow results. **Parameters:** - `domain_prefix` (optional): Filter by domain, e.g. `"com.acme"` - `intent_class` (optional): `"LOOKUP"`, `"QUERY"`, `"INGEST"`, or `"REVISE"` - `keyword` (optional): Free-text search acr
Data integrations for your lookups and summarization. Use whenever user wants info about his Google Calendar events (schedule), GMail emails, or Slack messages.
- 📄 _meta.json
- 📄 contract_reviewer.py
- 📄 SKILL.md
Review contracts and legal agreements (PDF, Word, images) for risks, unfair clauses, missing provisions, and key obligations using SoMark for accurate document parsing. Provides structured risk analysis with severity ratings. Requires SoMark API Key (SOMARK_API_KEY).
- 📁 .github/
- 📁 agents/
- 📁 docs/
- 📄 .coveragerc
- 📄 .gitignore
- 📄 CHANGELOG.md
Publication-grade medical prediction workflow with strict anti-data-leakage controls, phenotype-definition safeguards, lineage-based leakage detection, split-protocol verification, class-imbalance policy validation, hyperparameter-tuning isolation checks, falsification tests, and reproducibility gates. Use when building, reviewing, or debugging disease risk or prognosis models in EHR/claims/registry data, especially when target definitions, diagnosis codes, lab criteria, medications, temporal windows, and derived features can leak target information.
- 📁 .claude-plugin/
- 📁 commands/
- 📁 hooks/
- 📄 README.md
- 📄 SKILL.md
Strip sensitive EXIF metadata from images before publishing to the web. Activates when working with images for blog posts, websites, or public content, or when EXIF, metadata, GPS data, or image privacy is mentioned.
- 📁 assets/
- 📁 references/
- 📄 README.md
- 📄 SKILL.md
Design robust, scalable database schemas for SQL and NoSQL databases. Provides normalization guidelines, indexing strategies, migration patterns, constraint design, and performance optimization. Ensures data integrity, query performance, and maintainable data models.
Guide for adversarial machine learning: adversarial examples, data poisoning, model backdoors, and evasion attacks.
- 📁 .github/
- 📁 assets/
- 📄 .gitignore
- 📄 app-scaffolding.md
- 📄 business-workflows.md
Use when working with Auto.dev APIs, vehicle data, VIN decoding, car listings, vehicle photos, specs, recalls, payments, interest rates, taxes, OEM build data, plate-to-VIN, CLI commands, MCP tools, or SDK methods for any automotive data task
Use when you want your product to surface in AI-generated answers (ChatGPT, Perplexity, Gemini) — creates llms.txt, optimizes structured data, and configures AI crawler access for GEO.
Практическая инженерия диффузионных моделей: архитектуры, обучение, инференс, оптимизация памяти. Использовать при любых задачах с диффузионными моделями: проектирование или модификация архитектуры (UNet/DiT/Flow/Flux), выбор и настройка schedulers/samplers, дообучение (LoRA/DreamBooth/full fine-tune), оптимизация памяти (AMP/checkpointing/ZeRO/FSDP/quantization), замена или fusion текст-энкодеров (CLIP/Qwen), работа с Diffusers, отладка диффузионных пайплайнов, оценка качества (FID/CLIPScore/LPIPS), latent diffusion, VAE, guidance/CFG, rectified flow, Stable Diffusion, SDXL, Flux. Также применять при вопросах про GPU-память при обучении генеративных моделей, text-to-image пайплайны, ControlNet, multi-encoder fusion, WebDataset. --- # Diffusion Engineering Skill ## Быстрая ориентация Три инженерных решения, которые больше всего влияют на качество/скорость/стоимость: 1. **Где идёт диффузия** → пиксели (дорого) или латентное пространство (LDM/SD-семейство — практично) 2. **Backbone денойзера** → UNet (классика, проще) или Transformer/DiT/Flow (масштабируется лучше) 3. **Управление сэмплингом** → scheduler, число шагов, guidance_scale — часто дают больше, чем правка сети --- ## Reference files — читать по задаче | Тема | Файл | Когда читать | |---|---|---| | Архитектуры и data flow | `references/architectures.md` | DDPM/SDE/LDM/DiT/Flux/VAE/SDXL, схема пайплайна | | Schedulers и guidance | `references/samplers.md` | DDIM/Euler/Heun/DPM-Solver/PNDM, CFG, prediction_type | | Обучение и дообучение | `references/training.md` | Loss/цели, LoRA/DreamBooth/full FT, гиперпараметры | | Память и распределённость | `references/memory.md` | AMP, checkpointing, ZeRO, FSDP, quantization, FP8 | | Текст-энкодеры и данные | `references/encoders-data.md` | CLIP/Qwen/multi-encoder, токенизация, data pipeline | | Оценка и траблшутинг | `references/eval-debug.md` | FID/CLIPScore/LPIPS, типовые поломки и фиксы, лицензии | --- ## Быстрый чеклист «я строю/модифицирую diffusion» - [ ] **Backbo
