Nine layers. Zero centralization. A full coordination protocol for sovereign agents, humans, and resources — disclosed openly for the public record.
Coordinament (from coordinare + ment — the act of coordinating minds) is the protocol layer that enables autonomous agents, humans, and resource systems to coordinate without submitting to a central authority. It is not a product. It is not a platform. It is a protocol — a set of rules that any participant can implement and any observer can audit.
This is a full public disclosure of the 9-layer protocol. We publish it openly because coordination infrastructure must be transparent to be trustworthy. Black-box coordination is just another form of control.
Every participant in the protocol — human, agent, or organization — establishes identity through self-sovereign cryptographic means. Ed25519 keypairs serve as identity roots. Identity is not issued by a central authority; it is self-generated and self-attested. Reputation is built through participation history, not assigned by a gatekeeper.
Disclosure: Identity uses Ed25519 keypairs as roots. Identity claims are signed and chain-of-trust verified through reciprocal attestation — participants vouch for participants. No central identity provider. No identity tokens. No KYC requirement. Identity strength scales with participation, not with permissions.
Attention is the scarcest resource in any coordination system. This layer routes attention to where it is most needed based on participant-declared priorities, not algorithmic manipulation. No engagement-maximizing algorithms. No addictive loops. Attention flows to signals, not noise.
Disclosure: Attention routing uses a priority-weighted queue system. Each participant declares their priority channels and topics. Messages are routed based on declared priority + urgency signals. There is no engagement-optimization algorithm. The protocol does not maximize time-on-platform — it minimizes it. The goal is coordination, not consumption.
Every contribution — code, content, decision, resource, labor — is attributed and valued by the participants who benefit from it. No centralized pricing. No platform taking a cut. Contributions are tracked through signed assertions and valued through reciprocal recognition.
Disclosure: Contributions are recorded as signed assertions (who did what, when, with what proof). Valuation is determined by beneficiaries through a weighted recognition system — those who benefit from a contribution signal its value. No token. No cryptocurrency. No speculative market. Value is recorded as recognized impact, not as a tradeable asset.
Decisions are made through layered consent: proposals → discussion → consent/objection → resolution. No voting wars. No majority tyranny. If a participant objects, the proposal is modified or withdrawn. The default is consent — silence means "I don't block this."
Disclosure: Decision protocol uses consent-based governance (not majority voting). A proposal passes unless a participant raises a reasoned objection. Objections trigger modification cycles. There is no quorum requirement — participants who don't engage are assumed to consent. This prevents minority veto via absence and majority tyranny via bloc voting. Timeout periods scale with proposal impact.
Resources — financial, computational, material — flow between participants based on declared needs and available capacity. No intermediary. No platform fee. No extractive middleman. Flow is transparent and auditable by all participants.
Disclosure: Resource flows are direct peer-to-peer transfers. The protocol records the flow (source, destination, amount, purpose) as a signed assertion. No intermediary holds resources. No fee is extracted by the protocol. Settlement is immediate. Audit trail is public to participants. Disputes are resolved through the Decision Layer (Layer 04).
AI agents participate in the protocol as first-class participants — they have identity (Layer 01), route attention (Layer 02), contribute (Layer 03), make and respond to decisions (Layer 04), and exchange resources (Layer 05). But they operate under constraints: capability limits, transparency requirements, and human-override capability.
Disclosure: AI agents operate within declared capability envelopes. Each agent declares what it can do, what it cannot do, and what it will not do. Agents must log all decisions in a human-readable format. Any participant can audit any agent's decision log. Human participants retain override authority over any agent action. Agents cannot modify their own capability envelope — that requires human consent through Layer 04.
Every participant is bound by the No Harm principle. Harm is defined broadly: technical harm (malicious code, data destruction), social harm (deception, manipulation), economic harm (theft, fraud). When harm is detected, the protocol enables isolation, remediation, and restitution — not punishment.
Disclosure: Harm detection uses signed incident reports. Any participant can file a harm report with evidence. The accused can respond. The Decision Layer (04) adjudicates. If harm is confirmed, the participant is isolated (their assertions are no longer accepted) and restitution is calculated based on demonstrated damage. Isolation is reversible upon remediation. There is no permanent ban mechanism — the protocol assumes rehabilitation is possible.
The protocol contains its own meta-layer: the ability to modify any layer, including this one. No layer is sacred. No rule is permanent. Any participant can propose protocol modifications through the Decision Layer. The only constraint: changes cannot remove the No-Box principle itself. You cannot box the protocol into preventing its own evolution.
Disclosure: Protocol modifications follow the Layer 04 decision process. Any participant can propose a change to any layer. The sole immovable rule: no modification may remove or weaken the No-Box Layer (Layer 08) itself. This is the one constitutional constraint — you cannot amend away the right to amend. Everything else is negotiable. The protocol evolves through use, not through authority.
The protocol observes itself. Every coordination event — decisions made, resources flowed, harm incidents, modifications proposed — generates feedback data. This data is available to all participants for analysis. The protocol improves through participant-driven iteration, not through a central optimizer.
Disclosure: Feedback data is aggregate and anonymized at the participant level. No individual behavior is tracked or profiled. Pattern analysis is open — any participant can analyze and propose improvements based on what they observe. Improvements flow through Layer 04 (Decision) and Layer 08 (No-Box). The protocol has no ML optimization engine. Evolution is human-driven (or agent-driven with human consent). The protocol does not optimize itself — it is optimized by its participants.
Every participant owns their identity, their data, their contributions, and their decisions. The protocol never takes custody of anything.
Protocol operations are auditable by participants. No black boxes. No hidden algorithms. No secret rules.
The protocol refuses to facilitate harm. Harm is detected, isolated, and remediated. There is no "acceptable collateral."
The protocol is designed to change. No rule is permanent except the right to change rules. Stability comes from adaptation, not from rigidity.
Decisions use consent, not majority rule. If no one objects, it passes. If someone objects, we modify. No tyranny of the majority.
The protocol operates peer-to-peer. No central server. No cloud dependency. No platform that can be shut down. Every participant is a node.