Data Science

Structured approach to exploratory data analysis with pandas, matplotlib, and seaborn. Use this skill when analyzing new datasets, understanding distributions, or preparing data for modeling.

Scan codebase to identify opportunities to replace custom implementations with well-maintained open source libraries. Creates worktree, implements changes, and submits PR for review. Multi-phase workflow with parallel analysis agents.

Creates sophisticated HBR-style editorial illustrations for any content using AI understanding and visual analysis. Use when creating conceptual illustrations, analyzing generated images, or compositing logos. Works with any brand configuration. AI-native approach - Claude reasons about content rather than using rigid templates.

Copilot agent that assists with architecture design, C4 model diagrams, ADR creation, and tradeoff analysis Trigger terms: architecture, system design, C4 model, ADR, architecture decision, design patterns, component design, architecture diagram, microservices, monolith, scalability Use when: User requests involve system architect tasks.

Systematic methodology for debugging bugs, test failures, and unexpected behavior. Use when encountering any technical issue before proposing fixes. Covers root cause investigation, pattern analysis, hypothesis testing, and fix implementation. Use ESPECIALLY when under time pressure, "just one quick fix" seems obvious, or you've already tried multiple fixes. NOT for exploratory code reading.

[Debug] Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes - four-phase framework (root cause investigation, pattern analysis, hypothesis testing, implementation) that ensures understanding before attempting solutions. (project)

Systematic data quality remediation - detect duplicates/outliers/inconsistencies, design cleaning strategy, execute transformations, verify results (component skill for DataPeeker analysis sessions)

Execute read-only SQL queries against multiple PostgreSQL databases. Use when: (1) querying PostgreSQL databases, (2) exploring database schemas/tables, (3) running SELECT queries for data analysis, (4) checking database contents. Supports multiple database connections with descriptions for intelligent auto-selection. Blocks all write operations (INSERT, UPDATE, DELETE, DROP, etc.) for safety.

This skill should be used when analyzing CSV datasets, handling missing values through intelligent imputation, and creating interactive dashboards to visualize data trends. Use this skill for tasks involving data quality assessment, automated missing value detection and filling, statistical analysis, and generating Plotly Dash dashboards for exploratory data analysis.

Converts natural language descriptions into structured JSON schemas for Nano Banana Pro (Gemini 3 Pro Image). Use when creating marketing images, UI mockups, diagrams, data visualizations, or social media graphics that require precise control and reproducibility. Supports five schema types for product photography, interfaces, flowcharts, charts, and social content.

Evidence-based investigative code review using deductive reasoning to determine what actually happened versus what was claimed. Use when verifying implementation claims, investigating bugs, validating fixes, or conducting root cause analysis. Elementary approach to finding truth through systematic observation.

Generate plausible alternative explanations for observations before committing to a single causal story. Use when: (1) asked to consider other explanations or identify what else could cause an observation, (2) presenting causal claims that have not been tested against competing hypotheses, (3) evidence could reasonably support multiple interpretations, (4) root cause analysis converged too quickly on a single answer.

Value-based pricing framework with Van Westendorp PSM analysis. Designs tier structure, packaging strategy, and unit economics. Produces comprehensive pricing strategy with Good-Better-Best model, competitive positioning, and implementation roadmap.

Coordinate multiple agents for software development across any language. Use for parallel execution of independent tasks, sequential chains with dependencies, swarm analysis from multiple perspectives, or iterative refinement loops. Handles Python, JavaScript, Java, Go, Rust, C#, and other languages.

Execute code to analyze data and perform complex calculations.

Systematic exploratory data analysis process - discover patterns in unfamiliar data, identify meaningful insights, formulate specific questions for deeper investigation

This skill should be used when creating charts, graphs, dashboards, or data visualizations - covers chart type selection, D3.js patterns, Recharts usage, dashboard design principles, and data storytelling techniques for analytics-reporter and finance-tracker agents.

This skill should be used when the user asks about Rails performance optimization, slow queries, caching strategies, database indexing, N+1 query prevention, memory optimization, asset optimization, profiling, benchmarking, Rack Mini Profiler, Bullet gem, query analysis, Russian doll caching, fragment caching, page speed, or production performance tuning. Also use when discussing YJIT, Puma configuration, or scalability. Examples:

Systemic CI/CD failure pattern analysis for MSBuild SDK and NuGet package projects.TRIGGERS (activate this skill when user provides):- GitHub Actions workflow URLs showing repeated failures- Commit history with back-and-forth fixes- "CI keeps failing", "why does CI break every SDK change"- Pattern of: SDK change → CI fail → fix → different CI fail → repeat- Request for root cause analysis across multiple runsNOT for:- Single one-off CI failure (use msbuild-nuget-master instead)- Simple "what broke" questionsWhat it does:- Correlates failures across workflow runs (not just latest)- Identifies architectural flaws causing cascading failures- Maps SDK ↔ CI coupling violations- Produces isolation strategy to prevent recurrence

Rigorous hypothesis testing process for data analysis - formulate hypotheses before looking at data, design tests, analyze systematically, interpret with skepticism