Toolkit
Agentic Design Patterns
10 real-world agent designs that solve problems a single LLM call cannot. Each uses open source data, working code, and builds your intuition for when and how to use agentic architectures.
Use these when you have a specific business problem and want to see a full agent solution end-to-end.
Cyber Threat Intelligence Agent
Security teams drown in thousands of CVE feeds daily. This agent correlates vulnerabilities from the National Vulnerability Database with your specific infrastructure, cross-references MITRE ATT&CK techniques, assesses exploitability using EPSS scores, and generates a prioritized, actionable report.
Severe Weather Impact Agent
Businesses lose billions annually to weather disruptions they could have anticipated. This agent monitors NOAA forecasts and NWS alerts in real-time, maps severe weather events to business assets (warehouses, routes, facilities), assesses operational risk severity, and decides whether to trigger aut
Supply Chain Disruption Agent
Global supply chains have thousands of nodes spanning dozens of countries. When a disruption hits — a port closure, a semiconductor shortage, a geopolitical event — the cascading effects are unpredictable. This agent monitors supplier health indicators, detects anomalies in lead times and delivery r
School Teacher Assistant Agent
Teachers spend up to 50% of their time on lesson preparation, grading, and administrative tasks instead of teaching. This agent understands curriculum standards (Common Core, state standards), generates age-appropriate lesson plans, creates varied assessments aligned to learning objectives, grades s
Research Synthesis Agent
Researchers spend weeks conducting literature reviews across thousands of papers. This agent searches multiple academic databases (ArXiv, Semantic Scholar, CrossRef), fetches and analyzes abstracts, extracts key findings, detects contradictions between studies, and synthesizes everything into a cohe
Data Quality & Profiling Agent
An autonomous agent that profiles datasets, detects schema violations and outliers, suggests fixes, applies them iteratively, and validates the results. Every correction can reveal new issues that require re-analysis, making this a fundamentally iterative problem that static pipelines cannot solve.
Regulatory Compliance Agent
An autonomous agent that parses regulatory texts (EU AI Act, NIST AI RMF, GDPR), extracts actionable requirements, maps them against organizational policies, identifies compliance gaps, and suggests remediation steps. Regulations overlap and contradict each other — the agent must cross-reference mul
Open Data Investigator Agent
An autonomous agent that searches government data catalogs (data.gov, Census Bureau, BLS), downloads datasets, runs statistical analysis, detects anomalies, and generates narratives with visualizations. The exploratory workflow branches based on findings — discovering an anomaly in one dataset trigg
Incident Root Cause Agent
An autonomous agent that investigates production incidents by parsing structured and unstructured logs, correlating timestamps across services, checking dependent service health, forming hypotheses about root causes, and testing them against evidence. The agent follows the scientific method — observ
Personal Finance Advisor Agent
An autonomous agent that analyzes personal transaction data, categorizes spending, detects patterns and anomalies, compares against national benchmarks from the BLS Consumer Expenditure Survey, creates personalized budgets, projects savings trajectories, and generates actionable financial advice. Th