Panko Alerts — Food Safety Data

Searches across FDA, FSIS, and CDC recall databases simultaneously, aggregating results from 25+ government sources into a unified query interface. Implements a federated search pattern that normalizes heterogeneous government data schemas (FDA's OpenData format, FSIS XML feeds, CDC JSON endpoints) into consistent structured results with recall severity, affected products, and distribution scope.

Queries CDC foodborne illness outbreak surveillance data, enabling search by pathogen type, food vehicle, state, and date range. Implements epidemiological data normalization from CDC's FoodNet and PulseNet systems, returning outbreak case counts, affected demographics, implicated foods, and investigation status with temporal filtering for trend analysis.

Aggregates active food safety alerts from FDA, FSIS, CDC, and local health departments, providing a unified alert stream searchable by alert type, severity, and geography. Implements a polling-based aggregation pattern that normalizes alert formats across agencies (FDA warning letters, FSIS public health alerts, CDC health advisories, local health department notices) and returns structured alert metadata with actionable details.

Searches restaurant inspection records and compliance history from NYC, Chicago, SF, Seattle, and Boston health departments, returning inspection scores, violation details, remediation status, and historical trends. Implements a city-specific data normalization layer that maps disparate inspection scoring systems (NYC letter grades, Chicago risk categories, SF inspection scores) into comparable compliance metrics with violation categorization.

Searches curated food safety guides and best practices from FDA, FSIS, CDC, and local health departments, returning regulatory requirements, handling procedures, and compliance checklists. Implements a document retrieval pattern that indexes government food safety guidance documents and returns relevant excerpts with source attribution and regulatory authority, enabling programmatic access to food safety standards.

Exposes 5 food safety search and lookup tools via the Model Context Protocol (MCP) standard, implementing a schema-based function calling interface that allows MCP-compatible clients (Claude, other LLMs) to invoke tools with structured arguments and receive normalized JSON responses. Each tool is registered with a JSON Schema definition specifying input parameters, output structure, and tool documentation, enabling the client to understand tool capabilities and invoke them with type-safe arguments.