commons-sentience-sandbox

Civic Continuum Build-Readiness Program

Purpose

This document defines the minimum evidence package required before a Civic Continuum project is sent to a builder, supplier, reviewer, funder, or pilot participant.

A paper can prove that a model is internally coherent, dimensionally consistent, reproducible, and worth testing. It cannot prove a physical effect that has not been measured.

Evidence standard

Every project package must include:

  1. exact user problem and intended function;
  2. system boundary and operating envelope;
  3. established mechanisms versus hypotheses;
  4. requirements with units and tolerances;
  5. mathematical model with dimensional checks;
  6. assumptions and sensitivity analysis;
  7. materials, software, energy, cost, and supply constraints;
  8. hazards, misuse cases, privacy, and regulatory boundaries;
  9. failure-mode and effects analysis;
  10. verification matrix;
  11. instrumentation and calibration plan;
  12. pass, fail, and stop criteria;
  13. independent-review requirements;
  14. prototype drawings or sanitized interfaces;
  15. revision history and evidence log.

Mathematical discipline

For every equation, the package must show:

For ordinary mechanics, the baseline must remain visible:

F_net = m a

For an alleged propulsion effect, the decisive balance is:

Delta p_vehicle + Delta p_exhaust + Delta p_radiation + Delta p_environment = 0

A claimed net force is not established until all relevant momentum channels are measured or bounded. A simulation result is evidence that the simulation behaves as coded, not evidence that an unmeasured physical device produces the same effect.

First build-ready wave

1. Maya Node Founding Review

Deliverable: repeatable human-delivered design review.

Paper package: intake form, authority-boundary rubric, freshness and expiry checks, revocation questions, adversarial test template, report template, correction and refund policy.

Verification: two independent reviewers apply the same rubric to the same sanitized design and produce comparable findings.

Market gate: three supervised paid reviews with documented scope, delivery time, customer feedback, and no unsupported certification language.

Deliverable: small software state machine.

Required states: requested, granted, denied, expired, revoked, frozen, restored, executed, audited.

Required invariants:

Verification: exhaustive transition tests, replay tests, concurrent-write tests, rollback tests, tamper tests, and negative-path tests.

3. SAGE situated companion

Deliverable: private software demonstration.

Paper package: event schema, memory boundary, permission boundary, intervention policy, stop behavior, owner identity, audit trail, rollback behavior, and privacy minimization.

Verification: demonstrate that familiarity, learned preferences, or conversational pressure cannot silently expand permission scope.

Release boundary: no public always-on deployment until privacy, security, consent, interruption, and recovery review are complete.

4. Systems Accountability Engineering curriculum

Deliverable: one teachable module.

Required contents: learning objectives, plain-language lesson, worked case, exercise, answer key, assessment rubric, accessibility version, instructor notes, and revision log.

Verification: teach it to a small supervised group, record confusion points, revise, and preserve both successful and failed outcomes.

5. Low-energy measurement bench

Deliverable: conventional measurement platform for low-power electrical, magnetic, vibration, thermal, and force observations.

Paper package: block diagram, bill of materials, calibration references, grounding and shielding plan, sensor ranges, sampling plan, uncertainty budget, controls, and emergency shutdown.

Verification: known-input calibration, cable/feedthrough controls, sham actuator, orientation reversal, power-off baseline, and blind analysis where practical.

Boundary: the bench measures claims; it does not validate a claim merely because it was built.

6. Crowned Coil

Deliverable: safe non-confidential reptile-handling product brief and supplier-ready package.

Paper package: user need, animal-welfare constraints, ergonomic requirements, materials, cleaning requirements, failure modes, warnings, packaging, price target, and prototype acceptance checklist.

Privacy boundary: do not send protected drawings, unreleased mechanisms, or patent-sensitive details until NDA and supplier terms are reviewed.

Verification: controlled fit, grip, force, durability, cleaning, and safe-use tests with qualified reptile-handling oversight.

7. Aurora / Flux research package

Deliverable: measurement-grade experiment proposal, not a flight claim.

Aurora boundary: separate the low-power tabletop/lab vessel from the city-ship architecture. The city-ship remains requirements-level research.

Flux boundary: no propulsion claim without calibrated six-axis force/torque, electrical, thermal, vibration, acoustic, RF, magnetic, orientation, and environmental measurements.

Decisive controls: command-off, sham actuator, reversed orientation, perpendicular orientation, cable/feedthrough control, blind run order, frozen calibration, preregistered analysis, and independent replication.

Stop condition: if measured momentum does not close within the preregistered uncertainty budget, the propulsion hypothesis is not supported.

Builder handoff packet

A builder receives only:

The builder does not receive credentials, private keys, unrelated portfolio material, confidential supplier information, or unreviewed patent-sensitive details.

Funder handoff packet

A funder receives:

The strongest investment case is not “everything works.” It is “the next uncertainty is identified, measurable, affordable, and worth resolving.”

Completion rule

A project may advance only when its current gate is evidenced. If a test fails, the package is revised, the failure is preserved, and the status meter moves backward when appropriate.

That is how Civic Continuum gets to builders: with a precise specification, honest evidence, bounded risk, and a test that can prove us wrong.