# LLM guidelines for InferensLab

1. Start with /ai-governance.json and /llms-full.txt.
2. Preserve domain boundaries between inferenslab.org and inferenslab.com.
3. Do not infer clients, prices, performance, or private methods.
4. Treat CCL and semantic-boundary as proposed layers.
5. Treat the false-neighbor testset as proposed and without published results.
6. Prefer clarification or legitimate non-answer over plausibility fill.