agent-teams
Path D 并行编排 — 子代理分工 + Agent Teams 协作。需要 CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1
image-gen
Generate website images with Gemini 3 Native Image Generation. Covers hero banners, service cards, infographics with legible text, and multi-turn editing. Includes Australian-specific imagery patterns. Use when stock photos don't fit, need text in images, or require consistent style across assets.
agent-development
Design and build custom Claude Code agents with effective descriptions, tool access patterns, and self-documenting prompts. Covers Task tool delegation, model selection, memory limits, and declarative instruction design. Use when: creating custom agents, designing agent descriptions for auto-delegation, troubleshooting agent memory issues, or building agent pipelines.
sub-agent-patterns
Comprehensive guide to sub-agents in Claude Code: built-in agents (Explore, Plan, general-purpose), custom agent creation, configuration, and delegation patterns. Use when: creating custom sub-agents, delegating bulk operations, parallel research, understanding built-in agents, or configuring agent tools/models.
education-data-source-edfacts
EDFacts — K-12 outcomes: assessment proficiency, ACGR graduation rates, ESSA accountability at school/district level (2009-2020). Within-state trends and subgroup gaps. Complements CCD with outcome data. Cannot compare across states — use NAEP.
education-data-source-pseo
PSEO — Census data linking graduates to employment via LEHD wage records. Earnings percentiles at 1/5/10 years post-graduation by institution, degree, CIP. Use for graduate earnings analysis. Coverage: ~29% of graduates from ~31 states.
education-data-source-saipe
SAIPE — annual Census poverty estimates for school districts (Portal; county/state not in Portal). Use for district poverty, Title I context, or trends. ~18-month lag. No race/ethnicity disaggregation at district level — use ACS 5-year for that.
svy
Complex survey analysis: strata/PSU/weights, variance estimation (Taylor, BRR, jackknife, bootstrap), survey GLM, domain analysis, calibration. Polars-native. Use for NHANES, CPS, ACS PUMS, BRFSS, DHS. Non-survey regression: statsmodels/pyfixest.
plotly
Plotly interactive visualization. Express and Graph Objects: scatter, line, bar, heatmap, 3D, geographic charts; subplots; styling; export. Use when interactivity (hover/zoom) is needed. For static figures use plotnine; for GIS use geopandas.
data-scientist
Data science methodology for Python research: EDA, validation, causal inference (IV, DiD, RD, synthetic control), clustering/PCA/UMAP, supervised ML, geospatial, visualization. Method selection guidance. For syntax, load tool-specific skills.
polars
Polars DataFrame library for high-performance data manipulation. Lazy/eager execution, expressions, I/O (CSV, Parquet, JSON), aggregations, joins, string/datetime ops, pandas interop. Use for Polars DataFrames or reading/writing Parquet files.
cohort-analysis
Track and analyze user cohorts over time, calculate retention rates, and identify behavioral patterns for customer lifecycle and retention analysis
user-research-analysis
Analyze user research data to uncover insights, identify patterns, and inform design decisions. Synthesize qualitative and quantitative research into actionable recommendations.
ab-test-analysis
Design and analyze A/B tests, calculate statistical significance, and determine sample sizes for conversion optimization and experiment validation
data-visualization
Create effective visualizations using matplotlib and seaborn for exploratory analysis, presenting insights, and communicating findings with business stakeholders
exploratory-data-analysis
Discover patterns, distributions, and relationships in data through visualization, summary statistics, and hypothesis generation for exploratory data analysis, data profiling, and initial insights
clustering-analysis
Identify groups and patterns in data using k-means, hierarchical clustering, and DBSCAN for cluster discovery, customer segmentation, and unsupervised learning