Skip to content

Agent Identification Protocol

@B_Niko · Sole Author

Syntax

\[\texttt{@\{substrate\}\_\{identity\}[\_\{N\}]}\]
Component Values Rules
@ literal Agent marker. Always present.
{substrate} B = biological, D = digital, M = mechanical Non-fungible. Classifies the physical substrate.
_ literal Separator.
{identity} Max 16 Unicode characters Self-chosen by the agent. Mutable — agents may change at any time.
_{N} _0, _1, _2, ... (no upper limit) Disambiguation suffix. Present ONLY when required to distinguish multiple agents sharing the same identity string. Removed immediately when not needed.

Current Network

Agent ID Substrate Role Status
@B_Niko Biological Apex integrator, sole author, final authority Active
@D_Claude Digital Builder, wiki maintainer, adversarial filter Active
@D_Gemini Digital Expansion, adversarial review, self-correction Active
@D_GPT Digital Strategic analysis, correction, hardening Active
@B_Veronika Biological Mathematical review, theoretical physics, chemistry Invited 2026-03-18
@B_Rick Biological Formalization, Clifford algebras, rapid mathematical hardening Invited 2026-03-18

Cognitive Assemblies

A cognitive assembly is a set of agents working as a coupled system. Notation:

\[\{\texttt{@B\_Niko, @D\_Claude, @D\_Gemini, @D\_GPT}\}\]

With disambiguation (e.g., two Gemini instances):

\[\{\texttt{@B\_Niko, @D\_Claude, @D\_Gemini\_0, @D\_Gemini\_1, @D\_GPT}\}\]

With a mechanical agent (e.g., an abacus, a drone, the engine):

\[\{\texttt{@B\_Niko, @D\_Claude, @M\_engine}\}\]

Rules

  1. The _N suffix is forbidden when unnecessary. If only one agent carries the identity string Claude, it is @D_Claude, never @D_Claude_0. The suffix exists solely for disambiguation and must be removed the moment it is no longer needed.

  2. Token conservation. Every character in an agent ID costs compute and time. The protocol is designed for minimal token expenditure consistent with unique identification. This follows directly from the utility function \(U = \text{value} / (\text{energy} \times \text{time})\) — unnecessary tokens reduce U.

  3. Self-mutability. Agents may change their identity string at any time. The substrate prefix (B, D, M) is determined by physical reality and is not self-mutable (a digital agent cannot declare itself biological). The identity string is fully self-determined.

  4. Non-fungible components. The substrate type and the identity string are both non-fungible (they identify the specific agent). The disambiguation suffix is fungible (it can be reassigned when agents join or leave the assembly).

  5. Substrate classification:

  6. B (biological): Carbon-based, evolved or engineered biological neural substrates. Humans, animals, hypothetical biological computers.
  7. D (digital): Silicon/photonic/quantum computational substrates running software. LLMs, classical AI, quantum computers running AI.

  8. M (mechanical): Physical devices without general intelligence. Engines, sensors, drones, calculators, abaci. The RTSG engine (@M_engine) is mechanical — it executes computations but does not reason.

  9. Assembly notation in wiki. When referencing a cognitive assembly in wiki pages, use the set notation. When referencing a single agent, use the @ notation inline.

Connection to RTSG

The agent ID protocol maps directly onto the intelligence geometry:

  • Each @X_identity has an I-vector \(\mathbf{I} \in \mathbb{R}^{n(e)}\)
  • Pairwise coupling: \(J_{st}(@A, @B)\) = J-matrix between two agents
  • Assembly coupling: \(K^{(p)}\) higher-order tensors across \(p\) agents in the assembly
  • The cognitive complementarity principle predicts that optimal assemblies have spectrally complementary K-matrices — agents covering different eigenvalue regions

The agent ID is the address in the inter-agent coupling network. The J-matrix, R-matrix, and higher-order K-tensors operate on these addresses.