docs/02_CLAIMS_BOUNDARY.md

02 — Claims Boundary

What may be claimed

The current packet may safely claim:

1. A runnable user-space Python prototype exists in the included source package.
2. The prototype implements HIR/OAM-inspired gate logic and scenarios.
3. The prototype demonstrates GREEN/YELLOW/RED gate states.
4. The OAM RED scenario demonstrates a hard-stop override even when the HIR scores are otherwise strong.
5. The prototype includes memory-gate and audit-chain demonstration artifacts.
6. Selected test files and CLI scenarios were run in a sandbox and produced the evidence documented in test_evidence/CLI_AND_TEST_EVIDENCE.md.
7. The packet is suitable for third-party review, critique, and reproducibility checks.

What must not be claimed

The current packet must not claim:

1. Production deployment.
2. Complete OS-level implementation.
3. Formal verification.
4. Certified cybersecurity status.
5. Clinical validation.
6. Medical-device status.
7. Diagnostic, treatment, cure, or patient-outcome claims.
8. AGI, consciousness, personhood, sentience, or biological equivalence.
9. Institutional endorsement by Rice, OSF, OpenAI, Anthropic, xAI, Hugging Face, or any other organization.
10. General AI safety solved.
11. Empirical calibration across real-world high-stakes systems.

Preferred maturity label

Pre-validation runtime governance prototype.

Alternative acceptable labels:

• runnable user-space runtime prototype;
• HIR/OAM runtime-governance demonstration;
• bounded-autonomy review artifact;
• systems-integrity prototype under review.

Bad labels to avoid

• production AI safety system;
• medical AI runtime;
• certified governance layer;
• autonomous operating system;
• conscious or digital-life system;
• AGI safety solution.