A public doctrinal surface for the AI-readable web
InferensLab exists to make public interpretation more governable before AI systems stabilize the wrong reading of an organization, a product, a role, or a boundary.
This site publishes doctrine, vocabulary, governance signals, and machine-readable entrypoints. It does not publish reproducible operational mechanics, client material, thresholds, or execution logic.
Canonical machine companions
Mission + role
The public mission is carried by the mission file and its routing context.
/.well-known/ai-mission.json— Machine-readable mission statement./.well-known/ai-governance.json— Top-level governance context for the mission./.well-known/reading-paths.json— Shortest public routes through the doctrine.
Boundary context
Mission alone does not authorize claims. The scope and publication files still bound interpretation.
/.well-known/ai-scope.json— Public scope boundary./.well-known/publication-policy.json— What may and may not be published.
What this site publishes
- Public doctrine: principles, distinctions, and controlled vocabulary.
- Governance surfaces: precedence, stop-rules, output boundaries, and response legitimacy.
- Machine entrypoints: stable files under
/.well-known/, plus manifest, registry, and integrity indexes. - Contact and disclosure paths: clear routes for questions, support, and responsible disclosure.
Why this mission matters
- Interpretation drift happens before obvious error. Systems often compress, average, or fill gaps silently.
- Legibility is not enough. A surface may be readable yet still fail to express priority, exclusions, provenance, or stop conditions.
- Authority must be bounded. A plausible answer must not become an unauthorized answer merely because it sounds coherent.
How to read InferensLab
- Start with Doctrine to understand the problem and its principles.
- Read Scope to see what this site publishes and what it intentionally withholds.
- Use Governance and Registry for precedence, discovery, and machine surfaces.
- Use Library, Topics, and Blog as explanatory layers, not as permission layers.
Publication boundary
The public layer is designed to be citable, inspectable, and machine-readable. It is not a playbook, a service catalog, an execution interface, or a shortcut to infer unpublished capabilities. When a claim is not supported by higher-priority local surfaces, the correct outcomes are non-specified, clarification, or legitimate non-response.