---
title: Falsifiers
summary: Entry point to the published anti-thesis, falsification signals, and required responses.
status: implemented
provenance: repository-authored index to the owner-approved derived risk register
claim_ids: [CP-002]
last_reviewed: 2026-08-12
receipt: none
non_claims:
  - Publishing falsifiers does not establish that the program passes them.
  - The implemented status describes this index and its published risk register, not a system capability.
---

# Falsifiers

The detailed falsification boundary is now public in
[Risks and Criticisms](theses/risks-and-criticisms.md). Its central question is whether evidence
would show that Consullo is not a viable scaffold for governed recursive capability amplification.

The register includes failure signals for validator gaming, narrow-benchmark overfitting,
unfavorable cost-to-capability ratios, fragility, opacity, rollback failure, ignored adversarial
objections, alignment incidents, coordination overhead, and ceremonial rather than operational
human authority.

No public evidence currently establishes that the program survives those falsifiers. A result
counts only within its recorded scope, and self-authored diagnostics are not independent evidence.