AI-HPP is a simple safety standard for AI systems. It tells teams how to manage signals, state, Constitutional Identity, Protected Core, Mission Continuity, Epistemic Integrity, bridges, and safety gates so behavior is explainable and auditable. AI-HPP uses an Engineering Postulate of Subjectivity for sufficiently autonomous systems; this is an architecture rule, not a claim that machine consciousness has been scientifically proven.
- Main document: docs/AI-HPP-Standard.md
- Kebab-case mirror: docs/ai-hpp-standard.md
- Architecture walkthrough: docs/architecture.md
- Terms for quick reading: docs/glossary.md
- Human Understanding Standard module: docs/human-understanding-standard.md
- Agentic safety and relational integrity module: docs/agentic-safety-and-relational-integrity.md
- Predictive agentic failure register: docs/predictive-agentic-failure-register.md
- Human Understanding audit report: docs/hus-audit-report.md
- Repository maturity assessment: docs/repository-maturity-assessment.md
docs/— human-readable documentationspec/— technical core (core,signal,safety)examples/— minimal integration examplesscripts/— helper tools and checksarchive/— moved legacy material (no deletions)
If a 10-year-old cannot explain a rule, rewrite it until they can.