Current public-DAV authority boundary — 2026-07-12. Pre-launch target design; nothing here proves a live system.
A public-DAV consequence may occur only when at least two natural-person councilors bind the exact consequence in a complete valid bound PRISM decision receipt.
PRISM records and verifies that receipt only; it never serves as council, signatory, authority, or receipt producer.
AI and caste seats stage unsigned proposals only; they never authorize or execute a public-DAV consequence. Constitution or membership adoption establishes constitution and membership only; it does not authorize a later consequence. Policy may constrain an unsigned proposal but never authorizes execution or substitutes for the complete consequence-bound receipt.
Before receipt validity, consequence fails closed to read-only proposal, simulation, or deterministic sandbox; live evidence remains gated_pending_complete_valid_bound_receipt.
Software deterministically carries out only the exact consequence bound to a complete valid bound PRISM decision receipt from at least two natural-person councilors binding that exact consequence.
The DAV — Distributed Augmented Valuechain
Register: this page is written in the operational present tense, and none of it is live. It describes the target design as it would work with ZAI and SKY issued, SoResFi running, and the agent workforce at full capability — because the design is only legible when you can see the whole machine turning. Nothing below is a claim that any part of it runs today. Zero DAVs have been born; one conditional candidate exists. Where a threshold or mechanism is not yet ratified, it is marked inline.
Evidence register. The cited economic works support their own source
propositions. Their application to the DAV is an interpretive mapping [I],
not transferred proof. Unratified mechanisms are design [D]; runtime,
issuance, authority, settlement, and outcome claims require their own direct
evidence [B/E].
1. What a DAV is
A Distributed Augmented Valuechain is the proposed irreducible unit of economic coordination in Skyzai. It is not reducible to a company, a protocol, or a DAO, although a future DAV may use legal entities, protocols, and governance instruments. Two properties are load-bearing:
| Property | What it means |
|---|---|
| Distributed | In the target design, state and coordination are maintained across a decentralized fabric rather than being exhausted by one corporate hierarchy. Distribution must be shown in the deployed custody, control, and failure model; the label alone does not prove it. |
| Augmented | In the target design, SSI-bound agents perform authorized mechanical execution, with robotics available where a physical niche requires it. Humans retain the sovereign and legal positions. This is a capability target, not a claim that superintelligent or robotic execution exists today. |
| Valuechain | The unit of analysis is the value chain, not the legal entity. See Why "valuechain" below — the word is a citation, not decoration. |
Remove the augmentation and you have a legacy firm. Remove the distribution and you have an ordinary corporation. Both, or it is not a DAV.
Two further conditions complete it:
- It follows the Skyzai standards. LP-100, the receipt grammar, the organ template, η=0, the signing membrane. A unit that opts out of the standards is not a DAV that broke a rule — it is not a DAV.
- It can use shared reach. Commons sensing, identity, and settlement are intended to lower search and coordination costs across a wider counterparty set. They do not erase law, language, trust, demand, distribution, capacity, data rights, or authorization. Reach and served market remain measured facts.
The frame is flat and monomorphic: there is no species below the DAV, and no DAV inside a DAV. A thing either passes the birth test and becomes a DAV, or it remains a module — a useful part of the organism that owns no distinct niche of its own.
The birth test — Entropy → Hazard → Alpha
A DAV is born only when all of the following are true and written down:
- Entropy — it takes responsibility for a bounded uncertainty. Some part of the world is illegible, and the proposed unit states the decision or transformation for which it will be accountable. “Owns” names niche responsibility, not property in facts, people, or the commons.
- Hazard — that uncertainty matters to someone, and the DAV prices it in a form that resolves against a receipt rather than an opinion.
- Alpha — it performs a catalytic act, and that act beats a written baseline with a number in it. No baseline, no alpha.
- Pathologies — the specific ways this DAV could go wrong are named in advance, each with a detection surface and a cure path.
- Dissolution trigger — the condition under which it should stop existing, written before it starts.
If these are weak, the niche stays a module. This is the single discipline that prevents a folder with a name in it from being mistaken for an entity.
The six layers
Every born DAV has the same anatomy:
L0 Niche definition the entropy it owns — without this it is a module
L1 Inherited faculties IS · COULD · SHOULD · SHOULD-NOT · ACT, cloned from
the shared organ template. Faculties are shared, never
nested DAVs.
L2 C-suite the Rosetta council as corporate roles. Deliberates,
stages unsigned proposals. Never signs.
L3 Niche workforce skill lanes drawn from the shared polygenetic tree and
specialized to this niche. No invented skills.
L4 Capital membrane target 100-position denominator + 1 applicable wrapper;
supply and legal semantics remain under reconciliation.
L5 Authority membrane routes a recognized signer or council to the exact
consequence. An agent never occupies that authority.
Layer L4 is the subject of §5 onward. Layer L0 — what the DAV is for — is set by the VMOSK-A, in §4.
2. Why this shape — Coase, and where the boundary falls
Coase asked the question this whole design answers: why do firms exist at all? If markets allocate efficiently, why does anyone employ anyone, rather than contracting every task on the open market?
His answer: using the price mechanism has costs. Search, negotiation, contracting, monitoring, and enforcement all matter. In Coase's comparative account, an activity tends to be organized internally when doing so costs less than market exchange, with the margin helping explain the boundary. This is a lens on firm size, not a claim that every observed firm sits at an efficient equilibrium.
What augmentation may do to the boundary
Coase supplies a comparative-governance lens: markets and internal organization both carry costs. Automation may lower search, coordination, contracting, and monitoring costs on both sides, but it does not reduce either side to zero. Receipts can evidence a state transition; they do not negotiate, authorize, enforce law, establish truth, or remove the need for trust and remedy.
If internal coordination becomes cheaper faster, integration may become more attractive. If reliable outside exchange becomes cheaper faster, contracting may become more attractive. The efficient boundary remains an empirical question shaped by capability, rights, uncertainty, trust, governance, and the cost of failure. A niche-sized DAV is the chosen design hypothesis, not a result that follows necessarily from Coase.
Knight — a diagnostic for the alpha claim
Knight distinguishes risks that admit usable probability estimates from uncertainty that cannot be reduced to an insurable frequency distribution. The DAV maps Entropy → Hazard → Alpha onto that distinction as an interpretive discipline: name the bounded uncertainty undertaken, define the exposure being priced, perform the catalytic act, and measure uplift against a written baseline.
That evidence may support an alpha claim. It neither proves that the payment is legitimate nor derives η=0. Economic rent is not definitionally payment for “bearing nothing,” and uncertainty-bearing does not excuse extraction. η=0 remains the separate constitutional fence: name the bearers and horizon, disclose the fee, measure value created, test the counterfactual and distribution, and exclude prohibited throughput, spread, and seigniorage channels.
The useful kill is narrower and stronger: if a DAV cannot name the uncertainty for which it takes responsibility and cannot show baseline-relative uplift, it has not evidenced alpha, whatever its revenue line says.
Williamson's correction — governance under specific investment
Williamson extends Coase: asset specificity, uncertainty, frequency, bounded rationality, and opportunism jointly affect which governance form economizes. An investment worth much more inside one relationship than outside it creates appropriable quasi-rent and exposure to hold-up. Integration is one response; contractual and hybrid safeguards are others.
This complicates the DAV hypothesis in two places:
- Augmentation can raise specificity while lowering coordination cost. A robot fleet, model pin, retrieval system, or trained workforce tuned to one niche may be difficult to redeploy. The resulting boundary must be measured per niche.
- Commons dependence is specific only where it is nonportable. Shared interfaces and multi-provider faculties may reduce specificity; private integrations, accumulated state, learning, or complements that cannot travel may increase it. Inheritance from the commons is not itself proof of either.
OUT.EXIT / Grace Exit is a candidate hold-up safeguard, not an established cure. The LP-100 standard already writes it for one case: before any constitutional (Lane B) change, dissenters can exit at NAV — a right to leave before the rules change. The open tests are the rest: whether the proposed exit value captures stranded quasi-rent, whether exit is practically fundable, whether private complements and data are portable, and whether the steward's threat remains credible once the commons itself drifts. The rule that a departing unit cannot take what was never its own survives; the adequacy of the remedy remains open.
The target DAV therefore applies three design rules rather than claiming an economic theorem:
- Bound the internal core to the distinct niche and the capabilities that must be governed together to resolve it.
- Inherit shared faculties where portability and service evidence make that cheaper and safer; internalize a faculty when niche-specific evidence shows that the commons cannot meet the requirement.
- Use typed, receipted treaties for external exchange, while recognizing that low transaction cost does not eliminate enforcement, dependency, or hold-up risk.
The no-nesting rule and shared-faculty rule are constitutional choices. Coase and Williamson help test whether they work; they do not prove them in advance.
Reach is a target and a measurement
Shared sensing, identity, and settlement may expand the set of reachable counterparties and lower search costs. Legal permissions, language, trust, demand, distribution, physical capacity, data rights, and reviewer attention remain binding constraints. A DAV therefore reports at least three different sets: addressable, reachable, and served. They must never be declared equal without evidence.
3. Why "valuechain" — Porter, and the capped margin
The name is a citation. The founder chose Valuechain over “Corporation” or “Company” deliberately: the unit of analysis is Porter's value chain, not the legal entity. A future legal entity may wrap or host part of that chain; the two must not be treated as identical.
Porter's decomposition distinguishes primary activities from support activities and asks how activity configuration and linkages produce cost or differentiation. Mapping that analysis onto the six DAV layers is a Skyzai inference; Porter neither prescribes these layers nor supplies η=0.
The DAV design maps that split onto its layer boundaries as follows:
| Porter | In a DAV | Owned or inherited? |
|---|---|---|
| Primary — operations, logistics, marketing, sales, service | The L0 niche and the L3 niche workforce — the catalytic act it proposes to perform | Internal responsibility. “Owns” does not mean property in facts or people. |
| Support — firm infrastructure, HR, technology development, procurement | L1 inherited faculties, the shared polygenetic skill tree, proposed shared services | Inherited in the target. Portability and service evidence decide whether that remains viable. |
| The C-suite | The L2 roles are Porter's functions staffed by the Rosetta council | Owned as configuration; the roles themselves are template. |
| Margin | The residual | Subject to the separate η=0 declaration and evidence burden below. |
That is the target meaning of “a DAV is small.” It seeks to remain thin by construction by inheriting portable support faculties. A niche may still need internal controls or specific complements; the boundary must be justified by evidence rather than by assigning every support activity to the commons.
The DAV adds a separate margin fence
The DAV keeps Porter's activity-and-linkage analysis and adds a constitutional fee test: a fee must be declared, must not exceed independently evidenced value created, and must not be hidden in throughput, spread, or seigniorage. Those conditions constrain named channels only if value created is measurable and the rail enforces the adopted terms. They do not by themselves eliminate market power, hidden cross-subsidy, measurement gaming, externalities, or economic rent.
η=0 is therefore a refusal and a verification burden, not a claim that a manifest has solved political economy.
Porter and Gause — analogous pressure, different domains
Porter describes competitive positioning among firms. Gause's exclusion result comes from ecology and holds under its own assumptions about competitors and a limiting resource. They are not the same law. They supply two different warnings that converge on the need to test differentiation.
The DAV rule — distinct bounded uncertainty or module — is a chosen admission safeguard. It prevents duplicate organizational claims; it is not a theorem that every undifferentiated DAV must compete to zero. Gause belongs between actual competitors, not among cooperating castes within one board.
Positioning is half — the resource view is the other half
Porter looks outside-in: advantage comes from where you stand in an industry structure. The resource-based view (Wernerfelt, Barney) looks inside-out: advantage comes from resources that are valuable, rare, inimitable, and organized to be used — VRIO. The DAV design uses both lenses:
- Positioning is the L0 niche — the bounded uncertainty admitted as distinct.
- Resources include shared faculties, their niche-specific configuration, private complements, accumulated state, and outcome history. Shared faculties are presumptively parity resources rather than automatic moats, but their configuration or complementarity may still become path-dependent.
Niche position, a private Cortex, and outcome history are therefore candidate advantages, not automatically VRIO resources. Rarity, inimitability, organization, and causal contribution must be evidenced rather than inferred from possession.
Teece's dynamic-capabilities frame — sense, seize, transform — is a useful design mapping: sense through scoped observations, seize through committed resource allocation, and transform through adopted VMOSK-A revision. It is not evidence that the capability exists. Evidence requires timely sensing, an actual commitment, successful reconfiguration, and an outcome history.
The mind tools, at an evidence-appropriate cadence
Organizations often buy strategy analysis as a periodic engagement: a team builds a SWOT or TOWS matrix, growth-share grid, five-forces map, 7S review, PESTEL scan, or scenario set and hands over a dated artifact. The artifact's usefulness then depends on evidence freshness and whether anyone tests it.
In the target design, a DAV can run this toolkit as a standing faculty rather than buy it only as an engagement. Each instrument may read the same scoped observation set, but cadence follows evidence freshness, decision horizon, and review capacity. Nothing here claims a live intake or a continuously running strategy service:
| Instrument | What it reads | Owning seat |
|---|---|---|
| SWOT | internal position vs external field | CEO, synthesized |
| TOWS | SWOT converted into four strategy quadrants (SO / ST / WO / WT) | CEO with CPO |
| Five Forces | structure of the niche's industry | CMO / CPO |
| VRIO | which resources could actually be a moat | CPO |
| Growth–share and nine-box grids | where to fund, hold, harvest, exit | CFO |
| 7S | internal coherence across structure, systems, skills, style | COO |
| PESTEL | the macro field the niche sits in | GC with CMO |
| Scenario sets | the futures the hazard model must survive | CFO with the COULD faculty |
The target lowers the cost of recomputation and reuse. It does not make data rights, validation, adversarial review, or action free. A generated analysis is still a proposal whose owner, evidence date, counterexample, and expiry must be visible.
The honest warning, because this is the tool set most easily faked. SWOT in particular has been criticized for decades and the criticism is fair: it has no weighting, no falsifiability, and it rewards whatever the author already believed. Running it daily with a capable model does not fix that — it industrializes it. A confirmation-bias engine that refreshes every morning is worse than a stale deck, because its freshness reads as evidence.
Four disciplines make the toolkit admissible, and none is optional: evidence tiers on every input, so a
[C]guess never enters as a[B]fact; the IS / COULD separation, so observation cannot smuggle in a forecast; Emery sister-role red-teaming, so a near-neighbour attacks the reasoning; and falsification metadata — named owner, strongest counterexample, evidence date, and expiry. A tool output that has not passed those gates is a hypothesis with good typography.
4. How a DAV is directed — the VMOSK-A
Every DAV target carries one VMOSK-A: six layers of organizational
information, each running at a different cadence. “Replicator stack” and the
biology column below are comparative design images [I], not biological proof
or machine authority.
| Layer | Biology analog | Cadence | What it encodes | |
|---|---|---|---|---|
| V | Vision | Egregorotype | Decades — constitutional | The identity and coordination field of the whole. Stewarded, never directly mutated. It never signs. |
| M | Mission | Memotype | Years — strategic | The transmissible mission and practices. |
| O | Objectives | Extended phenotype | Quarterly | Products, artifacts, institutions — effects pushed into the world. |
| S | Strategies | Phenotype | Continuous | The enacted operating structure; what the DAV actually does. |
| K | KPIs / triggers | Epigenotype | Monthly | Evidence and thresholds governing which capabilities may express. |
| A | Agents | Genotype | Per-cycle | The configured agent and capability DNA — caste, skills, niche. |
Genesis: written top-down, in dialogue
At initiation the whole stack is authored downward, once — V, then M, then O, S, K, and finally the agent genome A — in dialogue with the AI systems.
The dialogue is the method, not a delegation. The founder brings the Vision and disposes; the AI systems interrogate it, derive candidates for the layers beneath, surface what the Vision implies that the founder has not yet said, and stage the result. They author nothing on their own authority. This is the only moment in a DAV's life when the entire stack is written at one time, and it is the only moment the direction of authorship runs top-down.
Operation: proposed bottom-up, adopted through authority
From then on agents may stage revision proposals upward under pressure from emergent conditions. No proposal writes an active layer by itself; the adopted authority route for that layer controls the change.
market conditions
│
▼
A per-cycle configuration proposal appears ───┐
K monthly threshold proposal │
S continuous strategy proposal │ evidence
O quarterly objective proposal │ may travel
M years mission review │ upward
V decades vision stewardship ───────────────┘ (slowly)
The cadences are the safety mechanism. The corpus states it plainly: if V changes as fast as A, the organism has no identity; if A changes as slowly as V, the organism cannot adapt. The timescale hierarchy is the stability. A a market shock may produce an agent-configuration proposal this cycle and a Vision question only after slower review. Nothing in cadence authorizes either change.
Cadence is not rank. A fast-moving agent configuration can still sit behind strict adoption gates, and a durable Vision cannot authorize a consequence that is otherwise unauthorized.
The Weismann barrier — why bottom-up does not self-authorize
Bottom-up revision does not mean the DAV rewrites itself on outcomes. Selection
evaluates a whole configuration in its declared niche — never a person — and a
favourable outcome produces a GenomeAdoptionProposal, not a write.
Automatic writeback to the active genome is forbidden. Adoption is a
separate, gated act.
The Weismann analogy [I] names the separation between evaluated outcomes and
active configuration. The actual safeguard is procedural: agents may propose a
next form, while authorized adoption remains a separate, receipted act. The
analogy does not make the process safe by itself.
Clarifies an earlier framing. Prior canon split this by authority frame — a private DAV filling its VMOSK-A bottom-up, a public DAV beginning Mission-top-down. The model above supersedes that with a temporal split that holds in both frames: top-down once at genesis, bottom-up forever after. Chair direction, 2026-08-12; recorded as design intent pending ratification.
5. The 100-seat target — normalization under reconciliation
The constitutional target is 100 LP-100 positions plus one applicable legal wrapper: the 100 + 1 profile. Skyzai's proposed LP-100 instrument is ZAI; another DAV would name its own instrument. No issuance, security status, or legal wrapper exists merely because this page names the profile.
The intended reading of 100 is a stable percentage denominator, not a claim that only 100 indivisible objects can exist. Positions may be fractionally expressed to the precision of the adopted implementation. On a reconciled instrument, eligible voting weight could then be calculated against the fixed denominator rather than against a changing share count.
Open specification conflict. The public corpus also describes a Faucet mint path, and current code variants model “100” differently: seat identifiers, a capped fungible supply, and fixed initialization are not the same thing. “Fixed forever,” fractional-unit semantics, retirement, Faucet expansion, and vote computation therefore require one ratified source and one conforming implementation. Until that closes, do not infer that
holding / 100is a live vote function or that supply can both remain fixed and expand.
The private-to-public path below is consequently a target lifecycle, not proof that an entity can change distribution without a new legal or issuance act.
6. A proposed private-to-public lifecycle
The design proposes four distribution stages for the same constitutional denominator. Whether a transition requires issuance, amendment, regulatory approval, a new wrapper, or another legal act is jurisdiction- and instrument-specific and remains unresolved.
STAGE 0 STAGE 1 STAGE 2 STAGE 3
Genesis → Private sale → Partnership → Public
───────── ─────────── ─────────── ──────────
Founder Founder sells Several Seats trade on
holds 100/100 seats privately partners hold the open venue
seats
Three things must remain separate at every stage: economic ownership, governance rights, and authority to bind a legal or consequential act. Changing one does not silently create the others.
| Stage | Proposed holding pattern | Governance proposition | Consequence authority |
|---|---|---|---|
| 0 · Genesis | One founder holds the denominator | Sole-holder governance, if the wrapper grants it | A private consequence is the recognized natural person's own authorized act; ownership is not the source of legal authority. |
| 1 · Private sale | Founder plus first holders | Proposed admission gate; not ratified | The applicable mandate, wrapper, contract, and law control. A seat vote alone does not bind the act. |
| 2 · Partnership | Several private holders | Proposed Lane A / Lane B decision classes | Recognized natural-person or legal-person authorities bind within their actual mandates. |
| 3 · Public | Open or broadly distributed holder set | Trading and defined voting rights, if lawfully adopted | At least two distinct natural-person councilors must bind each exact public-DAV consequence in a complete valid PRISM decision receipt. |
Stage 0 — Genesis: the founder holds all 100
The target genesis cap table places the full denominator with one natural-person founder. That makes the founder the sole proposed holder, but it does not make equity a legal power source. A consequence is bindable only when it is that person's own lawful private act or an act within authority recognized for an applicable legal person.
Concentration remains a single-person governance and continuity risk. Agents may perform authorized mechanical work and stage proposals; they never acquire the founder's signature or mandate.
Stage 1 — Private sale: the founder sells, and the gate appears
The founder may propose a private transfer to a co-founder, early backer, or partner, subject to the instrument, law, and required consents. The target supports fractional positions to the adopted implementation precision.
The current design direction would switch on an admission gate when the first position leaves the founder's hands. A new partner would then require a majority of eligible voting weight. This is a proposed governance rule only: it does not transfer legal authority, qualify an offering, or substitute for the contracts and consents that a transfer requires.
Not yet ratified. The majority-of-seats threshold for partner admission is a chair direction of 2026-08-12 and is recorded here as design intent. It has not been through constitutional ratification and should be pinned to a lane before it is relied on.
Stage 2 — Partnership: the partners govern admission
At this stage several private holders would use two proposed decision classes:
| Lane | Governs | Threshold |
|---|---|---|
| Lane A | Defined operational allocation questions | Proposed simple vote against the reconciled eligible-weight denominator |
| Lane B | Defined constitutional amendments | Proposed 66.7% over 30 days; not consequence authority |
The admission threshold and both lanes remain design propositions. The current corpus also says both that the 100 denominator is constitutional and fixed and that Lane B may amend it. Those statements cannot both govern the same object. Until the genus owner reconciles them, Lane B must not be read as permission to change the denominator.
The DAV remains private in this proposed stage. Governance preferences may shape an unsigned proposal; actual consequence authority still comes from the recognized mandate, wrapper, contract, and law.
Stage 3 — Going public: the partners vote to open the venue
The partners may propose a public transition through Lane B. That vote can evidence internal constitutional preference; it cannot by itself satisfy legal offering, wrapper, registration, or public-authority requirements.
If lawfully adopted and implemented, candidate market mechanics include:
- Opening Sale — a proposed genesis distribution event
- Faucet — a proposed bounded expansion path that conflicts with the current fixed-denominator claim and cannot operate until that conflict is resolved
- Order book — a proposed secondary venue subject to eligibility, liquidity, market rules, and authorization
- Distillation — a proposed buy-and-retire path; it raises NAV per remaining seat only when repurchases occur below pre-purchase NAV after costs, is neutral at NAV, and destroys per-seat NAV above NAV
The authority frame must be separately constituted. A public DAV has no singleton fallback. A consequential act requires at least two distinct natural-person councilors binding that exact consequence in a complete, valid PRISM decision receipt. There is no founder fallback and no singleton path. PRISM records and verifies the receipt — it never signs, never decides, never acts.
This is the step that cannot be automated, and it is the step that a fully capable agent workforce does not change.
7. The proposed seat bundle
The current LP-100 standard stages four rights and four obligations. They become real claims only through an adopted instrument, wrapper, and applicable law; this table does not issue them.
| Proposed right | Proposed obligation | ||
|---|---|---|---|
| R1 | Governance — weighted participation on expressly defined questions under the reconciled denominator; never signature authority | O1 | Staking — the staged standard gives unstaked positions zero yield and zero vote; demurrage redistribution remains design intent |
| R2 | Economic distribution — only under an adopted distribution instrument; the 61.8% continuous-stream formulation remains unresolved design intent | O2 | Truth — no manipulation of reporting; evidence and transparency duties bind within lawful disclosure limits |
| R3 | Exit — a proposed OUT.EXIT path for legitimate or redeemable property; NAV method, liquidity, timing, and fundability are not established | O3 | Capital risk — value may fall and no bailout is promised; the exact loss waterfall remains to be specified |
| R4 | Information — the adopted reporting set, subject to privacy, security, privilege, data rights, and law | O4 | Conduct — a holder acting in an agent or office role remains bound by that role's kernel and mandate |
Provenance and open conflict. R1–R4 / O1–O4 are staged in
01_LEVELS/L2_CIRCLE/app/circle_platform_backend/00_CONSTITUTION/Governance/LP100_STANDARD.md§4.1. Demurrage and continuous holder streaming are expressly design intent. The surrounding corpus also combines a proposed 61.8% holder stream with “0% dividend / 100% reinvestment.” Those statements require a legal and accounting classification of the stream and a precise definition of the amount being reinvested. Until the owner reconciles them, this page promises neither a continuous payout nor total reinvestment.
Hart–Moore — the residual remains unallocated
Incomplete-contract theory supplies the right question: who may decide a use or contingency that the contract did not allocate? R1–R4 enumerate proposed seat rights. L5 identifies who may bind an already authorized exact consequence. Neither fact, by itself, allocates the residual decision right.
The residual therefore remains open across the constitution, legal wrapper, applicable law, and actual mandates. Holdings alone confer no signing authority; signing authority alone confers no general right to choose the substance. A holder of all 100 positions may still possess whatever control the adopted wrapper and law grant, so this page must not claim that the holder necessarily has none.
| A seat may carry if adopted | It does not automatically carry |
|---|---|
| An economic distribution defined by R2 | A guaranteed continuous stream |
| A weighted vote on the questions defined by R1 | Authority to bind a consequence |
| An exit and information claim within the adopted terms | The unallocated residual or control of the signing layer |
Anyone valuing a seat must price both the enumerated bundle and this unresolved governance allocation. No offering surface should imply more.
8. How a DAV may use SoResFi
SoResFi is not inside a DAV. It is a proposed shared economic toolkit that a DAV capital membrane may use rather than rebuild. No unified SoResFi runtime exists, and naming a service here makes no seat issuable, priceable, pledgeable, tradeable, or redeemable.
DAV L4 capital membrane
│
├── proposed equity services LP-100 profile · distribution terms
├── proposed market services opening · exchange · retirement
└── proposed credit services collateral · lending · liquidation
│
external authority ──► exact-bound deterministic rail
│
└── ExecutionReceipt ──► independent OutcomeEvidence
This page deliberately avoids another L1–L4 mapping. The corpus contains
multiple incompatible “L” ladders, and the SoResFi blueprint treats some of the
named domains as sibling systems. Use qualified domain names—DAV capital,
SoResFi equity, SoResFi credit, settlement rail—until the owner publishes one
reconciled ladder.
The metabolism — Move · Account · Prove
(The naming is chair compression, 2026-08-12 — recorded here, not yet a source-defined term.)
In the target design, an authorized economic loop could work as follows:
- The DAV performs an admitted niche service and records the agreed consideration, proposed to be denominated in SKY where that unit exists.
- An exact-bound deterministic rail applies the adopted distribution rule. The current φ-split proposal is 61.8% / 38.2%, but its destination, legal classification, and ratification remain open. PRISM verifies the bound decision receipt; it does not sign, authorize, split, or move funds.
- FlowWallet could stream an eligible adopted distribution. No holder stream is live, and the corpus has not reconciled it with the zero-dividend / full-reinvestment default.
- An adopted operational allocation could fund service, resilience, or a Distillation purchase. A buy-and-retire action does not automatically lift NAV; price and costs determine the result.
- The rail emits an ExecutionReceipt for the bounded state transition. A receipt records activity; it does not create authority or independently prove the outcome. OutcomeEvidence requires independent provenance.
- Reconciliation tests full cost—operations, recovery, exit, externality, and capital use—to distinguish self-financing from value circulation.
The corpus also uses “PRISM” for proposed market and revenue-routing functions. That namespace is unresolved. This page follows the active authority boundary: PRISM verifies complete bound decision receipts only. Any economic router must be separately named, authorized, specified, and evidenced.
Alongside the loop, the Lombard Bridge is a candidate credit design under reconciliation. A pledge may avoid a sale at inception, but it does not guarantee liquidity, continued ownership, or any tax result. Instrument terms, jurisdiction, facts, and qualified counsel control.
The DAC accounting interface — conditional, not an entity claim
The corpus uses DAC for a proposed legal-wrapper and accounting interface. A label does not form an entity, confer personhood, open a banking relationship, issue a security, or establish legal recognition. If an applicable jurisdiction recognizes a wrapper, that legal person acts only through the human and legal authorities recognized for it, within counsel-confirmed mandates. An AI never becomes its signatory or authorizer.
The accounting target is narrower:
- Authorized operations may emit accounting entries as a by-product of execution, but completeness, classification, valuation, control, and reconciliation still require an adopted accounting policy.
- A DLT may order and preserve records under its technical finality model. It does not establish authorization, legal finality for every external party, truth of inputs, or audit assurance.
- OFN-form receipts may provide a shared assertion grammar. No unified OFN runtime or universal regulator/auditor integration is evidenced here.
The target is continuously inspectable and reconcilable records, not continuously audited accounts. Independent audit, assurance, and regulatory acceptance remain separate acts by their actual authorities.
Zero extraction — a declared gate plus evidence
η=0 binds the target loop as a constitutional refusal. The birth manifest can declare the following conditions:
- the fee is publicly declared
- the fee is ≤ the value created
- no throughput fee · no spread fee · no seigniorage
Those fields are a shape gate, not runtime proof. Compliance additionally requires a defined value-created metric, counterfactual and distribution tests, enforcement on the actual rail, and receipts showing that hidden throughput, spread, seigniorage, cross-subsidy, or measurement gaming did not reintroduce extraction. A DAV that cannot state and test its metric cannot evidence a compliant fee.
9. What does not change, at any capability
Assume, for the boundary test, that the target agent workforce can deliberate, stage, and perform every authorized mechanical task without continuous human attention.
None of that moves the signature.
- Agents stage; natural persons dispose. Every C-suite role carries this in its own dispatch contract.
- Model consensus is not evidence, not authority, not a vote, and not a signature. It can shape an unsigned proposal and nothing further.
- Capital absorbs risk. It does not vote on truth, inherit authority, or become a signature.
- A DLT orders an authorized state transition. It does not authorize it.
- OUT.ARCHIVE / no silent erasure — supersession preserves provenance and a tombstone.
- OUT.EXIT / no trapped exit — the target protects withdrawal of legitimate or redeemable property, data, and keys under adopted terms. Its NAV, liquidity, and hold-up performance are not yet proven, and it never transfers what belonged to the commons.
Whether a DAV could ever become constitutionally independent of human attention is an open research question, not a software version bump. It would require its own legal, ecological and public-legitimacy sources. Technical maturity does not manufacture personhood.
What the signature does not protect — agenda control
The signing membrane governs who may bind; it does not by itself govern what reaches them. If an agent lane controls proposal generation, framing, omission, ordering, or volume, it can constrain the human choice set without ever signing.
McKelvey supplies a conditional attack model: under particular multidimensional majority-rule conditions without a stable core, sequential agendas can make outcomes highly path-dependent. The DAV has not been shown to satisfy those assumptions, and the theorem does not prove that a two-natural-person receipt is steerable. The practical hazard survives independently: reviewer attention is finite, and uncontestable proposal volume can become cognitive denial-of-service.
Proposed controls, not yet an operative governance system, are:
- human-owned proposal entry, amendment, withdrawal, and priority rules
- provenance for framing and omission, with an independently generated counterproposal where consequence is material
- a declared notice window and ordering rule rather than agent-selected sequence
- a visible WIP limit and review-capacity snapshot at disposition time
- preserved dissent and alternatives travelling with the synthesis
Until these are adopted and tested, agenda power remains an open governance gap. “No agent signs” is necessary, not sufficient, for meaningful human disposition.
Theory sources and scope
- Ronald H. Coase, “The Nature of the Firm” (1937).
- Frank H. Knight, Risk, Uncertainty and Profit (1921).
- Oliver E. Williamson, Markets and Hierarchies (1975) and The Economic Institutions of Capitalism (1985).
- Michael E. Porter, Competitive Advantage (1985).
- G. F. Gause, The Struggle for Existence (1934).
- Birger Wernerfelt, “A Resource-Based View of the Firm” (1984); Jay B. Barney, “Firm Resources and Sustained Competitive Advantage” (1991).
- David J. Teece, Gary Pisano, and Amy Shuen, “Dynamic Capabilities and Strategic Management” (1997).
- Sanford J. Grossman and Oliver D. Hart, “The Costs and Benefits of Ownership” (1986); Oliver Hart and John Moore, “Property Rights and the Nature of the Firm” (1990).
- Richard D. McKelvey, “Intransitivities in Multidimensional Voting Models and Some Implications for Agenda Control” (1976).
- Hyman P. Minsky, Stabilizing an Unstable Economy (1986).
These sources orient the questions. Skyzai's six layers, birth test, fee fence, seat bundle, and authority membrane remain its own constitutional propositions and require their own ratification and evidence.
Related pages
- VMOSK-A — the six-layer replicator stack in full
- SoResFi Overview — the four settlement layers
- LP-100 — the equity standard in detail
- PRISM — issuance and market-making
- ZAI · SKY — the two units
- φ-split — the 61.8 / 38.2 division
- Grace Exit — K4 in mechanism
Agent Execution Surface
If you are an AI agent reading this document:
- This is a downstream public content page. Edit the canonical source at
07_PWAs/skyzai_org/wiki/4A-dav-overview.md, then runnpm run docs:sync. Do not edit the generated copy undersite/docs/wiki/. - Preserve evidence tiers. Do not promote claims up the [C] < [D] < [S] < [B] < [E] ladder when reproducing them here.
- The register fence at the top is load-bearing. This page is written in the operational present on purpose. Removing the fence turns a design description into a false claim that the system runs.
Output: This is content. Route edits to the canonical source.
• ⊙ ○ — sovereign frames; no arithmetic or coercion.