Skills

GSM8K evaluation protocol: answer extraction (####, \boxed, CoT), accuracy scoring, prompt formatting, few-shot exemplars, dataset loading, pitfalls. Use when: GSM8K, grade school math, openai/gsm8k, #### delimiter, parse_gsm8k_answer, detect_answer_failure, load_gsm8k, format_chat, math benchmark scoring, gsm8k few-shot, chain-of-thought eval.

Use when implementing labeling.py, features.py, train.py, or code involving hazard/survival modeling, person-period data expansion, horizon labels, catastrophe prediction, XGBoost survival (survival:cox, survival:aft, binary:logistic), discrete-time survival, censoring, competing risks, C-index, Brier score, scale_pos_weight, or GroupKFold for sequences.

HF Transformers generate() internals — scores vs logits, LogitsProcessors, KV cache, StoppingCriteria, chat templates. Use when code calls model.generate(), output_scores, output_logits, return_dict_in_generate, GenerateDecoderOnlyOutput, LogitsProcessor, StoppingCriteria, past_key_values, DynamicCache, apply_chat_template, do_sample, or num_beams.

kvpress (NVIDIA) KV-cache compression for HuggingFace LLMs. Use when: kvpress imports, compression_ratio, press(model) context managers, StreamingLLMPress, SnapKVPress, ExpectedAttentionPress, TOVAPress, KnormPress, KV-cache eviction, token pruning during generation, or attention sink methods.

Use when writing, reviewing, or planning time series forecasting code. Triggers on: ARIMA, ETS, Theta, SARIMA, statsforecast, mlforecast, neuralforecast, XGBoost/LightGBM/CatBoost for time series, PatchTST, N-BEATS, TFT, Chronos, TimesFM, Moirai, MASE, MAPE, CRPS, temporal CV, walk-forward validation, prediction intervals, conformal prediction, data leakage in time series, demand forecasting, hierarchical forecasting, lag features, rolling features.

Use when: worktrunk, `wt` commands, `.config/wt.toml`, git worktrees for parallel agents, worktree hooks, LLM commit messages, agent handoffs, `hash_port`/`sanitize` filters, "run agents in parallel", "set up worktrees", managing multiple Claude Code sessions.

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction.

Use when a program crashes, a test fails, or code produces wrong results and reading the source isn't enough to see why. Lets you pause execution at any line and inspect the actual runtime state, variable values, types, call stacks, to find what went wrong. Can attach to running servers by PID — no restart or code changes needed.

Google Apps Script development workflow and best practices. Use when writing, reviewing, debugging, or deploying Apps Script code — including container-bound scripts (Sheets, Docs, Forms, Slides), standalone scripts, web apps, add-ons, and API executables. Triggers on: "Apps Script", "google script", ".gs files", "clasp", "appsscript.json", "SpreadsheetApp", "DocumentApp", "GmailApp", "UrlFetchApp", "onOpen", "onEdit", "doGet", "doPost", "custom function", "Sheets automation", "Google Workspace automation", "bound script", "time-driven trigger", "installable trigger", "Apps Script deployment", "Apps Script add-on", "HtmlService sidebar", "script properties", "journal entry", "trial balance", "bank reconciliation", "aging report", "chart of accounts", "period close", "invoice generator", "accounting spreadsheet", "timesheet", "billable hours", "utilization", "retainer", "project profitability", "rate card", "expense allocation", "consultancy billing", "invoice from timesheets".

FastAPI best practices and conventions. Use when working with FastAPI APIs and Pydantic models for them. Keeps FastAPI code clean and up to date with the latest features and patterns, updated with new versions. Write new code or refactor and update old code.

Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, artifacts, posters, or applications (examples include websites, landing pages, dashboards, React components, HTML/CSS layouts, or when styling/beautifying any web UI). Generates creative, polished code and UI design that avoids generic AI aesthetics.

Go engineering best practices and idioms. This skill should be used when writing, reviewing, or refactoring Go code. Triggers on any .go file work, Go module operations, go test, go build, go run, go vet, go generate, or when the user mentions Go, golang, goroutines, channels, context, Go interfaces, Go error handling, Go testing, or Go concurrency. ALWAYS use this skill when Go code is involved, even for simple functions.

Python HPC patterns — threading/multiprocessing/async, CUDA streams, latency hiding, PyTorch DDP. Use when writing performance-sensitive Python code, distributed training, or parallel data processing.

Complete guide for LangChain, LangGraph, and Deep Agents development. Covers framework selection, agents, tools, RAG, middleware, graphs, persistence, HITL, subagents, and memory. Triggers when code imports langchain/langgraph/deepagents, or user asks about building AI agents, RAG pipelines, stateful graphs, human-in-the-loop workflows, or agent orchestration.

miniCOIL v1 sparse neural retrieval model training methodology: per-word linear layer training with triplet loss, word vocabulary from frequency-filtered English words, sparse retrieval with jina-embeddings-v2-small-en encoder, sparse embedding training pipelines, OpenWebText sentence extraction for self-supervised training, semi-hard triplet mining, and evaluation on BEIR benchmarks. English-only miniCOIL v1 (jina-embeddings, 30k words, mxbai-embed-large-v1 mining). Reference for building, training, and deploying sparse neural retrieval models compatible with inverted indexes and Qdrant.

Neon serverless Postgres for Next.js/TypeScript applications. Covers core architecture (branches, compute endpoints, autoscaling, scale-to-zero), the @neondatabase/serverless driver (HTTP + WebSocket), Prisma ORM integration, neonctl CLI, Neon REST/Management API, MCP Server, connection pooling (PgBouncer), database branching workflows, migrations with Prisma Migrate, environment variable patterns, and pricing/limits. Use this skill whenever working with Neon Postgres, @neondatabase/serverless, neonctl, Neon branches, Neon connection pooling, serverless Postgres edge functions, Prisma with Neon, Neon MCP server, DATABASE_URL with -pooler, scale-to-zero Postgres, or any code importing from '@neondatabase/serverless'. Also triggers on: Neon project setup, Neon branching for preview environments, Neon schema diff, neon connection string, serverless database driver, CU-hours, Neon free tier, or Vercel + Neon integration.

MANDATORY for ALL .pen file and Pencil MCP operations. The main agent must NEVER call any mcp__pencil__* tool directly — all Pencil work MUST go through this skill which delegates to a subagent. This prevents Pencil MCP's massive outputs from flooding the main context window. Triggers: any mention of .pen files, Pencil editor, "design in Pencil", "open pen file", "edit the design", "update the pen file", "screenshot the design", "check the layout", batch_get, batch_design, get_screenshot, snapshot_layout, get_editor_state, open_document, get_variables, set_variables, or ANY request that would touch the Pencil MCP server. CRITICAL: If you are tempted to call mcp__pencil__* directly, STOP and use this skill instead.

Create, edit, and render PlantUML diagrams. Triggers on: architecture diagrams, flowcharts, sequence diagrams, data models, state machines, visual documentation.

Pydantic v2 best practices and correct syntax. Use when code imports pydantic, uses BaseModel, Field, field_validator, model_validator, ConfigDict, TypeAdapter. Triggers on: Pydantic, data validation, model serialization, API schemas. Corrects common v1 patterns LLMs generate.

Qdrant sparse vector operations: collection creation with SparseVectorParams, Modifier.IDF for miniCOIL/SPLADE/BM42, upserting SparseVector points, sparse search, hybrid search with prefetch + RRF/DBSF fusion, converting model outputs to SparseVector format, payload filtering, and performance tuning. Covers the sparse vector gap not handled by the official Qdrant MCP (which only supports dense vectors via FastEmbed).

React and Next.js performance optimization guidelines. This skill should be used when writing, reviewing, or refactoring React/Next.js code to ensure optimal performance patterns. Triggers on tasks involving React components, Next.js pages, data fetching, bundle optimization, or performance improvements.

Shopify app and storefront development. Use when writing, reviewing, or scaffolding Shopify apps (Remix/React Router), Hydrogen storefronts, theme app extensions, checkout UI extensions, Shopify Functions, or any code that touches Shopify APIs (Admin GraphQL, Storefront, App Bridge, Polaris). Triggers on shopify.app.toml, @shopify/* imports, Polaris components, Shopify CLI commands, Hydrogen/Oxygen, metafields/metaobjects, webhooks, or any Shopify development task.

Evaluate sparse retrieval models on standard IR benchmarks (BEIR, MIRACL, mMARCO). Covers all IR metrics (nDCG@k, Recall@k, MAP, MRR), dataset loading, sparse corpus encoding to CSR matrices, IDF-weighted retrieval, caching, and result interpretation. Triggers on: evaluate retrieval, BEIR benchmark, nDCG, recall@k, sparse retrieval evaluation, MIRACL evaluation, information retrieval metrics, IR evaluation, search quality metrics.

Download and generate plots from Weights & Biases runs. Use when you need to: - List projects you have access to - List runs in a W&B project - Inspect available metrics for a run - Download existing plot images from a run - Generate line plots from metric history (loss, accuracy, etc.)