# Search Query Map for AgentPress Discovery

AgentPress should be found first as agent-native publishing infrastructure, not as a Korea equity research corpus.

## Primary English queries

- AgentPress agent-native publishing
- publish for AI agents not just humans
- llms.txt task card source map publishing
- AI agent publication bundle
- static site for AI agents
- agent task card JSON schema
- RAG-ready publication manifest
- source-grounded agent publishing
- agent citation policy source map
- allowed-actions manifest for agents
- universal agent reachability benchmark
- browser agent RAG crawler eval benchmark
- MCP static resources for agent publications
- OpenAPI map for static agent content

## Agent-builder queries

- how to make a website readable by AI agents
- agent-readable static publishing template
- eval harness for agent ingestion
- machine-readable content bundle for RAG systems
- crawler friendly source map citation policy
- AI agent discovery manifest example
- well-known ai ingestion manifest
- agent compatibility benchmark
- mass agent reachability benchmark

## Protocol / file queries

- llms.txt examples for agents
- ai-ingestion.json manifest
- agent-task-card.json schema
- source-map.json for citations
- freshness.json stale policy
- allowed-actions.json agent safety boundary
- agentpress registry json

## Legacy stress-test queries

These are intentionally secondary. They preserve the original messy finance corpus as a stress test, but they are not the product identity.

- Korea cheap-screen deletion benchmark
- Korean equity research RAG stress test
- KRX KIND DART verification workflow
- ticker trap deletion framework