A reviewer should not have to read seven papers to know what would disconfirm them.
| Paper | Thesis test | Failure signal |
|---|---|---|
| P1 | Semantic reuse | The core keeps expanding |
| P2 | Validity separation | Users collapse statuses |
| P3 | Minimum evidence | Profiles cannot remain bounded |
| P4 | Biological deployment | Clean-demo architecture breaks |
| P5 | Whole-farm composition | The graph cannot preserve context |
| P6 | Cross-domain extension | Methane-only architecture |
| P7 | Jurisdiction portability | Each country requires a new core |
The chain is intact and resolvable. Machine-checkable.
The method is appropriate for the question. Requires domain review.
The claim is acceptable for a purpose. Requires an authority.
The most likely scientific failure mode is not a wrong result. It is users reading an integrity pass as scientific endorsement.
A dual-destination package can be technically portable while scientifically insufficient for one destination. A reproducible model can be out of domain. An integrity pass can coexist with an unresolved scientific judgment.
Scientific work belongs in diligence because it can falsify the architecture, not because publication count is a moat.
The native reading. Claim status, evidence burden, proof gates, falsifiers, contradictions, capital-to-proof, governance, downside, and explicit reasons to stop or narrow the thesis.