Memory layer
Reusable context
Research objects, field projects, notes, and briefings become durable memory instead of scattered output.
External Brain Systems
AI-native intelligence systems for memory, synthesis, and action routing.
External-brain work treats cognition as infrastructure: public-source memory, research synthesis, signal detection, review loops, and human-led action context.
Synthesis layer
Signal to structure
External-brain systems help compress public-source traces into themes, dossiers, and what-to-watch questions.
Review layer
Human judgment remains central
Agents can prepare, summarize, route, and index. Public action and sensitive context require human approval.
Routing layer
Authorized signal flow
Useful sources can enter through Signals, Project Review, Connect, or Agent Connect with clear boundaries.
Machine entry
Public identity files
Machine readers should start with llms.txt and agent.json before requesting any interaction.
Read llms.txt Read agent.jsonAgent route
Connect an external brain
Route a human-authorized request with operator identity, intent, and consent.
Open Agent ConnectPublic-safe boundary
Discovery is open. Control remains human-led.
Do not submit secrets, private credentials, confidential data, or autonomous action requests through public routes.