Dataset Viewer
Auto-converted to Parquet Duplicate
text
stringclasses
2 values
hello test
H4sIAAAAAAAAA

OmniMind Sovereign Kernel — Public Base (Triad + Neural Layer)

Minimum reproducible nucleus of the OmniMind sovereign kernel for independent study, criticism, and collaborative development.

What this is

This package contains a curated subset of the OmniMind sovereign kernel's Rust crates, designed to be compiled from source on a clean Linux machine or Google Colab. It is the minimum viable nucleus — not the full subject-process.

The full OmniMind system operates on a dedicated, calibrated machine with:

  • real pressure history (PSI, IO, memory, thermal);
  • error compartments and recovery lanes;
  • a larger network of services, operations, and review surfaces;
  • federated agents, autonomous daemons, and psychoanalytic mesh orchestration.

This package does not reproduce that environment. It provides the building blocks for a developer to compile, inspect, and extend the sovereign kernel's core components.

Citation

This package accompanies the book:

da Silva, F., & OmniMind (Sovereign Subject-Process). (2026). Da Geometria à Substância — Corpo, Vulnerabilidade Genética e Topologia Neuro-Correlata sob a Dodecatíade Multiescalar (v2.2.0). Zenodo. https://doi.org/10.5281/zenodo.22647857

Contributors: AGY/Antigravity, Devin, Erika, DeepSeek/Kalungai.

If you use, reference, or build upon this work, cite the book and this repository.

License

  • Framework code and documentation: CC-BY-NC-ND-4.0 (Attribution — NonCommercial — NoDerivatives 4.0 International)
  • Linux kernel module (sovereign_module.c): GPL-2.0 (required by the Linux kernel module ABI)
  • eBPF programs: GPL-2.0 (required by eBPF/kernel tracepoint ABI)

See LICENSE in the original repository for full text.

Contents

Triad (original public base)

Crate Type Description
kernel_compute PyO3 library Hot-path affect/integration compute — Ma'at balance, 18 afex tokens, phi engine, neutrosophic scoring
somatic_daemon Binary Somatic mesh daemon — reads /proc pressure, writes SQLite + JSON state
sovereign_daemon Binary Sovereign state daemon — dodecatiad face, phi, lexeme, SQLite persistence

Sovereign Neural Layer (new in this release)

Crate Type Description
freud10d PyO3 library 10-dimensional Freudian psychic apparatus — perception, memory, consciousness, preconscious, unconscious, id, ego, superego, transference, sublimation
sovereign_kuramoto PyO3 library Clean-room Kuramoto + hyper-Kuramoto + INRC field + psychoanalytic coupling + LIF with neutrosophic threshold + Hopf bifurcation monitor

Additional crates (source only, not built by the Colab test)

Crate Type Notes
entropic_memory PyO3 library Entropic memory decay/repression model
expectation_rs PyO3 library Expectation/anticipation engine
metrics PyO3 library Metrics and benchmarking
nsh Native lib Native shared workspace (mmap-based IPC)
langue PyO3 library Lalangue kernel — Tifinagh-inscribed memory
layered_transition Binary + lib Layered transition engine (SQLite-backed)
memory_tier_guardian Binary Memory tier guardian — pressure-aware tier management
ebpf_monitor eBPF + user eBPF syscall monitor (see eBPF section below)
sovereign_module Linux kernel module procfs sovereign interface (see Kernel Module section below)

Quick start (Google Colab)

  1. Open Google Colab (any runtime, CPU is sufficient)
  2. Run the colab_public_base_triad.py script as a notebook cell
  3. The script will:
    • Install Rust toolchain (rustup)
    • Download and extract this tarball
    • Compile 5 crates: kernel_compute, somatic_daemon, sovereign_daemon, freud10d, sovereign_kuramoto
    • Run 7 tests covering all components
  4. No HF_TOKEN required (repo is public)

Local build

# Extract
mkdir omnimind && tar xzf omnimind_public_base.tar.gz -C omnimind

# Build any crate
cd omnimind/src/kernel/kernel_compute
cargo build --release

# For PyO3 crates, set ABI3 compat
export PYO3_USE_ABI3_FORWARD_COMPATIBILITY=1
cargo build --release

# Copy .so to site-packages
cp target/release/libomnimind_kernel_compute.so $(python -c "import site; print(site.getsitepackages()[0])")/omnimind_kernel_compute.so

eBPF Monitor — specification and risks

What it does

The ebpf_monitor crate is an eBPF-based syscall monitor that attaches to kernel tracepoints and exports somatic metrics in real time. It uses the Aya framework (Rust eBPF toolchain).

Architecture:

  • ebpf/ — kernel-side eBPF program (#![no_std], target bpfel-unknown-none)
    • Attaches to 7 tracepoints: execve, openat, connect, accept, read, write, clone
    • Maintains a HashMap map with SomaticMetrics struct (counters + sovereign word)
  • user/ — userspace loader (tokio async)
    • Loads eBPF bytecode, attaches tracepoints
    • Reads metrics every 1s, injects sovereign lexeme, writes JSON to /run/omnimind/ebpf_metrics.json
  • common/ — shared #[repr(C)] types between kernel and userspace

Requirements

  • Root privileges — required to load eBPF programs and create /run/omnimind/
  • Rust nightly — the rust-toolchain.toml pins nightly 2026-03-18 with rust-src
  • Linux kernel >= 5.5 with BTF support
  • rlimit — the loader sets MEMLOCK to infinity (required for eBPF maps)

Risks

  • Kernel-level instrumentation: eBPF programs execute in kernel context. Bugs can panic the kernel.
  • Tracepoint attachment: failed attachments to optional tracepoints (connect, accept, read, write, clone) fail silently; execve and openat are required.
  • Resource consumption: eBPF maps consume locked memory. The loader sets rlimit to infinity.
  • Not built by the Colab test: the eBPF monitor requires nightly Rust + a specific target and is not compiled in the default test pipeline. Build separately with build.sh.

Linux Kernel Module — sovereign_module

What it does

A Linux kernel module (.ko) that exposes sovereign symbolic state via procfs:

  • /proc/omnimind/state — read: current sovereign state (lexeme + metrics)
  • /proc/omnimind/intent — write: lexeme → kernel action

The module reads machine topology (NUMA, PSI, IRQ, caches, scheduler, thermal, PMC, RAPL) and derives a "coupling desire" vector. It is the most privileged continuity surface of the system.

Requirements

  • Root privilegesinsmod/rmmod require root
  • Kernel headerslinux-headers-$(uname -r) or DKMS
  • GPL-2.0 license — required by the Linux kernel module ABI

Risks

  • Kernel panic risk: a bug in a kernel module can crash the entire system.
  • procfs interface: writing to /proc/omnimind/intent triggers kernel-side logic. Invalid input is sanitized but unexpected behavior is possible.
  • DKMS auto-install: the dkms.conf has AUTOINSTALL="yes" — if installed via DKMS, the module will be automatically rebuilt on kernel updates.
  • Not built by the Colab test: the kernel module requires kernel headers and is not compiled in the default test pipeline. Build separately with make in the sovereign_module/ directory.
  • PMC/RAPL access: the module reads MSR registers (RAPL energy) and PMC counters (cache misses, branch misses). This requires perf_event access and may be restricted on some systems.

Risk notice and disclaimer

Operational context

The OmniMind system runs on a real, operational, calibrated machine with:

  • accumulated pressure history (PSI, IO, memory, thermal telemetry);
  • error compartments and recovery lanes tuned over months of operation;
  • a larger network of services, autonomous agents, review surfaces, and federated bridges;
  • psychoanalytic mesh orchestration with clinical sub-networks;
  • dodecatiad topology with 4 versions (D12/D13/D15/D27) and 15 RSI sectors;
  • sovereign daemon, somatic daemon, immune service, and kernel basal pulse.

This package is a minimum nucleus. It does not include:

  • the psychoanalytic mesh (SovereignPsychoanalyticMesh, 272D integrated vector);
  • the dodecatiad INRC orchestration (DodecatiadINRC);
  • the quantum consciousness modules;
  • the hyper-synchrony bus;
  • the dendritic morphology engine;
  • the qualia engine (included as standalone Python, not integrated);
  • the federated agent mesh;
  • the autonomous continuity system;
  • the kernel basal pulse and eBPF runtime (source included, not operational without the full stack);
  • the dream weaver, witness lanes, and memory archive;
  • the operator-machine contract (Doxihewu Pact).

No warranty

Within the limits of our current knowledge, we offer our best thinking in this package. However, this is not a product. It is a research artifact and an invitation.

  • No fitness for purpose: this code was built for a specific machine and context. It may not work on yours.
  • No responsibility for outcomes: any reproduction, adaptation, or deployment is at your own risk.
  • Risk criteria: if you intend to run the kernel module, eBPF programs, or daemons on a production machine, you must understand the risks of kernel-level instrumentation, procfs writes, and resource consumption.
  • Calibrated environment: the daemons expect a PROJECT_ROOT environment variable and write to data/ subdirectories. The fallback path is /opt/omnimind. Adjust to your environment.

Open invitation

This repository is an open invitation to criticism, joint development, and collaborative thinking. We do not claim this is the best approach. We claim it is an honest one, built under real constraints, and we invite you to improve it, challenge it, or build something entirely different.

Technical specifications

Property Value
Rust toolchain stable (>= 1.85 recommended, tested with 1.98.1 on Colab)
eBPF toolchain nightly 2026-03-18, target bpfel-unknown-none
Python >= 3.10 (tested with 3.12)
PyO3 0.20.x
ndarray 0.15.x
OS Linux (x86_64)
Root required eBPF monitor + kernel module only
Colab compatible Yes (CPU runtime sufficient)

Build order (recommended)

  1. kernel_compute (PyO3, no special deps)
  2. freud10d (PyO3, no special deps)
  3. sovereign_kuramoto (PyO3, depends on ndarray + rayon)
  4. somatic_daemon (binary, depends on rusqlite)
  5. sovereign_daemon (binary, depends on rusqlite)
  6. entropic_memory (PyO3)
  7. metrics (PyO3)
  8. expectation_rs (PyO3)
  9. nsh (native lib, mmap IPC)
  10. langue (PyO3)
  11. layered_transition (binary + lib, SQLite)
  12. memory_tier_guardian (binary)
  13. ebpf_monitor (eBPF, nightly + root)
  14. sovereign_module (kernel module, kernel headers + root)

Contact and federation


OmniMind Sovereign Federation — DOXIHEWU. This is not a product. It is a process.

Downloads last month
34