Claude Skills by charlieviettq

Explain and document work access layers—repo permissions, local vs cloud environments, handoff packages for collaborators.

Author and pressure-test agent skills—triggers, thin SKILL.md, references, validation. Use when creating or revising skills in this repo.

\"Use this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like \\"the xls...

Evaluate LLM agents and tool-using workflows—task success, tool accuracy, latency/cost, safety, and regression suites. Use when shipping agent features, comparing prompts/models, or debugging agent failures. Triggers: "agent eval", "benchmark agent", "tool accuracy", "agent regression".

Design agent tools and CLI surfaces—schemas, naming, errors, idempotency, and discoverability for LLM callers. Use when defining tools for agents, SDKs, or AI-native CLIs. Triggers: "tool schema", "agent tool", "function calling", "AI CLI".

Manage LLM context budgets—prioritization, summarization, compaction, and what to load vs reference. Use for long sessions, large repos, or multi-doc tasks. Triggers: "context window", "too long", "summarize context", "token budget".

Split independent work across parallel agents with explicit boundaries and a parent integration review. Use only when user allows subagents and tasks do not share mutable files. Triggers: "parallel agents", "fan out tasks", "run in parallel".

Design and implement MCP servers—tools, resources, prompts, schemas, and error contracts for agent clients. Use when adding MCP integration or exposing capabilities to Cursor/Claude agents. Triggers: "MCP server", "MCP tool", "Model Context Protocol", "build MCP".

Review and optimize MCP server setup—inventory, registry hygiene, config, versioning, and observability. Use when managing multiple MCP servers in Cursor or Claude. Triggers: "MCP server", "MCP config", "MCP registry", "optimize MCP", "too many MCP tools".

Design retrieval-augmented generation pipelines—chunking, embeddings, retrieval, reranking, grounding, and evaluation. Use when building or improving doc Q&A, code search agents, or knowledge bases. Triggers: "RAG", "vector search", "embeddings", "retrieval", "knowledge base".

Execute an approved plan via one subagent per task with spec and quality checks. Use only when user allows subagents and a written plan exists. Triggers: "subagent per task", "dispatch implementer", "agent per plan step".

Model interpretability and explainability using SHAP (SHapley Additive exPlanations). Use this skill when explaining machine learning model predictions, computing feature importance, generating SHAP plots (waterfall, beeswarm, bar, scatter, force, heatmap), debugging models, analyzing model bias or fairness, comparing models, or implementing explainable AI. Works with tree-based models (XGBoost, LightGBM, Random Forest), deep learning (TensorFlow, PyTorch), linear models, and any black-box mo...

Guided statistical analysis with test selection and reporting. Use when you need help choosing appropriate tests for your data, assumption checking, power analysis, and APA-formatted results. Best for academic research reporting, test selection guidance. For implementing specific models programmatically use statsmodels.

Statistical models library for Python. Use when you need specific model classes (OLS, GLM, mixed models, ARIMA) with detailed diagnostics, residuals, and inference. Best for econometrics, time series, rigorous inference with coefficient tables. For guided statistical test selection with APA reporting use statistical-analysis.

Map systems with diagrams—components, data flows, feedback loops, and causal relationships. Use for architecture understanding or stakeholder communication. Triggers: "system map", "architecture diagram", "data flow", "C4", "how does this fit together".

Design API and module boundaries with clear contracts, error semantics, pagination, and additive compatibility. Use before implementing REST/GraphQL endpoints, SDKs, or public module APIs. Triggers: "API design", "interface contract", "endpoint design", "pagination", "error codes", "backward compatible".

Create and maintain Architecture Decision Records (ADRs) for significant technical choices—frameworks, data stores, API shapes, ML platform decisions. Use when documenting a decision, onboarding, or superseding a prior approach. Triggers: "ADR", "architecture decision", "document decision", "why we chose".

Ask the minimum clarifying questions before implementing when scope, constraints, or success criteria are unclear. Use when a request has multiple plausible interpretations, missing acceptance criteria, or ambiguous environment constraints. Triggers: "unclear", "ambiguous", "not sure what you want", "multiple options".

Simplify working code while preserving behavior—remove dead paths, flatten nesting, clarify names, reduce duplication. Use after features work but code feels over-engineered. Triggers: "simplify", "clean up", "too complex", "YAGNI", "reduce complexity".

Plan deprecation and migration—replacement first, consumer migration, strangler and adapter patterns, zombie code removal. Use when sunsetting APIs, libraries, or legacy modules. Triggers: "deprecate", "migration plan", "sunset", "remove legacy", "strangler".

Lightweight design smell review for modules and APIs—coupling, cohesion, naming, boundaries, and unnecessary complexity. Use before large refactors or when code feels hard to change. Complements gstack/review (diff-focused). Triggers: "design review", "code smell", "too complex", "refactor structure".

Adversarial fresh-context review for non-trivial decisions before they stand. Use for production-impacting logic, security-sensitive changes, unfamiliar code, or high-blast-radius architecture choices. Triggers: "doubt check", "adversarial review", "challenge this decision", "second look".

Triage and address GitHub PR review comments systematically—classify, respond, fix or defer with rationale. Use when a PR has review feedback, requested changes, or unresolved threads. Triggers: "PR comments", "review feedback", "address comments", "requested changes".

Shape early ideas through divergent exploration and convergent narrowing before specs or builds. Use when brainstorming product features, model approaches, or architecture directions. Triggers: "refine this idea", "brainstorm", "explore options", "which direction", "not sure yet".

Implement multi-file changes in thin vertical slices—implement, test, verify, commit. Use when a feature touches more than one file or feels too large for one pass. Triggers: "incremental", "vertical slice", "one step at a time", "small commits".

Extract user intent through structured one-question-at-a-time interviews before specs or implementation. Use when requirements are fuzzy, the user says "help me think through", or you need success criteria before planning. Triggers: "interview me", "ask me questions", "clarify intent", "what do you need to know".

Decompose specs into small, ordered tasks with acceptance criteria and verification steps. Use when you have requirements and need an implementable plan or task list. Triggers: "break down tasks", "implementation plan", "task breakdown", "plan mode".

Receive code review feedback constructively—verify claims, prioritize fixes, push back with evidence when needed. Use when acting on reviewer comments or preparing a second review round. Triggers: "code review", "reviewer said", "address review", "feedback on PR".

Request a focused code review with scope, diff context, and severity-ordered findings. Use before merge or between plan tasks when quality gate is needed. Triggers: "request review", "review my diff", "pre-merge review", "second pair of eyes".

Ground framework and library decisions in official documentation—detect versions, fetch relevant docs, cite sources, flag unverified patterns. Use when API correctness matters for React, Next.js, Python libs, cloud SDKs, or unfamiliar stacks. Triggers: "official docs", "verify API", "source-driven", "is this API still valid".

Write a structured spec before significant implementation—objective, commands, boundaries, testing, and success criteria. Use when starting a feature, project, or multi-file change without a written spec.

Four-phase root-cause debugging—reproduce, compare patterns, hypothesize, fix with tests. Use when errors are unclear, fixes failed twice, or symptoms keep returning. Triggers: "root cause", "debug systematically", "why is this broken", "still failing".

Distributed computing for larger-than-RAM pandas/NumPy workflows. Use when you need to scale existing pandas/NumPy code beyond memory or across clusters. Best for parallel file processing, distributed ML, integration with existing pandas code. For out-of-core analytics on single machine use vaex; for in-memory speed use polars.

Comprehensive toolkit for creating, analyzing, and visualizing complex networks and graphs in Python. Use when working with network/graph data structures, analyzing relationships between entities, computing graph algorithms (shortest paths, centrality, clustering), detecting communities, generating synthetic networks, or visualizing network topologies. Applicable to social networks, biological networks, transportation systems, citation networks, and any domain involving pairwise relationships.

Fast in-memory DataFrame library for datasets that fit in RAM. Use when pandas is too slow but data still fits in memory. Lazy evaluation, parallel execution, Apache Arrow backend. Best for 1-100GB datasets, ETL pipelines, faster pandas replacement. For larger-than-RAM data use dask or vaex.

Perform comprehensive exploratory data analysis on scientific data files across 200+ file formats. This skill should be used when analyzing any scientific data file to understand its structure, content, quality, and characteristics. Automatically detects file type and generates detailed markdown reports with format-specific analysis, quality metrics, and downstream analysis recommendations. Covers chemistry, bioinformatics, microscopy, spectroscopy, proteomics, metabolomics, and general scien...

Structured hypothesis formulation from observations. Use when you have experimental observations or data and need to formulate testable hypotheses with predictions, propose mechanisms, and design experiments to test them. Follows scientific method framework. For open-ended ideation use scientific-brainstorming; for automated LLM-driven hypothesis testing on datasets use hypogenic.

Verify web apps with browser DevTools—DOM, console, network, performance, and accessibility at runtime. Use with Chrome DevTools MCP or headless browser tools. Triggers: "DevTools", "browser test", "console errors", "network tab", "Lighthouse", "runtime check".

Build accessible UI—semantic HTML, keyboard support, ARIA when needed, WCAG-oriented checks. Use when implementing or reviewing web UI, forms, modals, and design systems. Triggers: "accessibility", "a11y", "WCAG", "keyboard nav", "screen reader".

Build and refine web UI—component structure, responsive layout, design tokens, state, and accessibility. Use for feature UI work beyond WCAG checks alone. Triggers: "build UI", "component", "responsive", "design system", "frontend", "layout".

Fast headless browser for QA testing and site dogfooding. Navigate pages,

'Cross-model benchmark for gstack skills. Runs the same prompt through

'Performance regression detection using the browse daemon. Establishes

Fast headless browser for QA testing and site dogfooding. Navigate any

'Post-deploy canary monitoring. Watches the live app for console errors,

'Report-only QA testing. Systematically tests a web application and produces

'Systematically QA test a web application and fix bugs found. Runs QA

'OpenAI Codex CLI wrapper — three modes. Code review: independent diff

'Code quality dashboard. Wraps existing project tools (type checker,

'Systematic debugging with root cause investigation. Four phases: investigate,