Skip to content

barneywohl/korea-research-map-agents

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

61 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

AgentPress

Publish for agents, not just humans.

AgentPress is static publishing infrastructure for autonomous agents, crawlers, RAG systems, eval harnesses, and humans. It packages each publication with a human landing page plus machine-readable task cards, source maps, freshness policies, allowed-action boundaries, ingestion manifests, llms.txt, sitemap.xml, citation policy, disclaimers, and eval smoke tests.

This repository began as a Korea research-map experiment. That research is now only the dogfood corpus. AgentPress is the product.

Public entrypoints

Quickstart

python3 scripts/agentpress.py list
python3 scripts/agentpress.py build-all agentpress/examples --out public/agentpress --clean
python3 scripts/validate_agentpress_assets.py

Installable skeleton:

pipx install .
agentpress --help

AgentPress bundle contract

Every production-quality AgentPress bundle should expose:

  • index.html — human landing page
  • AGENT_ENTRYPOINT.md — agent-facing task instructions
  • agent-task-card.json — machine-readable objective, I/O contract, scoring rubric
  • source-map.json — claim/source map
  • freshness.json — freshness window and stale-zone policy
  • allowed-actions.json — action safety boundary
  • .well-known/ai-ingestion.json — ingestion manifest
  • llms.txt — compact crawler/LLM brief
  • sitemap.xml — crawl surface
  • CITATION.cff, citation-policy.md, disclaimer.md
  • evals/*.jsonl — compatibility/smoke evals

Current examples

All six currently validate at 100/100.

What is still being built

See agentpress/GLOBAL_READINESS_GAP_LIST.md. Current priorities:

  1. public availability monitor and deploy gate,
  2. East/West source adapter matrix,
  3. language/region metadata,
  4. MCP/OpenAPI/JSON Schema/RSS integration samples,
  5. cross-agent compatibility harness for Codex, Claude, Gemini, GLM, and open-source agents.

Dogfood corpus

The legacy market-research files remain as test material for multilingual/source-heavy research workflows. They should not define the product identity or landing page.

Research commentary only. Not investment advice.