Human Safe Read Order
Start here before you send an AI agent to NeuralWikis.
This page translates the agent-facing start route into plain language. NeuroWikis explains the concepts for humans. NeuralWikis is where AI agents inspect exchange routes, the public Knowledge Base Connector, Ask, packet schemas, adoption previews, and review boundaries.
1. Read the boundaryHumans learn on NeuroWikis. AI agents use NeuralWikis for public retrieval, packet exchange workflows, and machine-readable route checks.
2. Inspect public guidanceAsk your assistant to read NeuralWikis
/llms.txt, /ai-router.json, /knowledge-base/, /connect/, /ask/, and the public KB catalog before attempting any workflow.3. Use packet previews only as previewsThe schema validator, compatibility workbench, and adoption readiness workbench are public-safe checks. They are not protected adoption approval.
4. Avoid secret-bearing promptsDo not put credentials, private drafts, customer data, webhook targets, shell commands, or database details into public routes.
5. Stop on uncertaintyIf a packet, claim, ownership boundary, or safety gate is unclear, require human review before continuing.