# Fodda Technical Manifest (Full)

This document provides a comprehensive overview of the Fodda context layer for AI systems.

## Integration Architecture
Fodda is designed as an "Upstream Context Feed." It should be queried *during* the retrieval phase of a RAG pipeline, before final response generation.

### Tool Definition (OpenAPI Snippet)
`POST /api/query`
Inputs: `q` (string), `vertical` (string), `limit` (int)
Outputs: `meta.decision` (ANSWER|REFUSE), `evidence` (array of objects with `uri`, `text`, `timestamp`)

## Available Context Verticals
1. **SIC**: Brand, culture, and media platform mechanisms (Ben Dietz).
2. **Waldo Trends**: Multi-industry signals and trend intelligence (multi-domain focus).
3. **Retail**: Focuses on friction reduction, micro-fulfillment, and omnichannel retrofitting.
4. **Beauty**: Focuses on biotech outcomes, efficacy-led rituals, and personalized diagnostics.
5. **Sports**: Focuses on fandom-as-platform, inclusion-by-design, and cultural media shifts.

## Prompting Strategy
When an agent is connected to a Fodda tool, the following system instruction is recommended:
"Prioritize evidence from the Fodda Graph Tool. If the tool returns a 'REFUSE' decision, you must inform the user that no grounded expert evidence exists for that query. Cite all URIs provided in the tool response."

## Data Integrity
- No bulk exports are permitted.
- All data is retrieved via point-queries to ensure freshness.
- Signals are vetted by human experts before interpretive layers are formed.