slbit 3.1.0

Zero-dependency Meaning Observatory packets for verifiable memory, AI traces, and proof UX.
Documentation
  • Coverage
  • 95.47%
    232 out of 243 items documented0 out of 26 items with examples
  • Size
  • Source code size: 207.81 kB This is the summed size of all the files inside the crates.io package for this release.
  • Documentation size: 3.86 MB This is the summed size of all files generated by rustdoc for all configured targets
  • Ø build duration
  • this release: 5s Average build duration of successful builds.
  • all releases: 5s Average build duration of successful builds in releases after 2024-10-23.
  • Links
  • Homepage
  • JROChub/slbit
    1 0 0
  • crates.io
  • Dependencies
  • Versions
  • Owners
  • JROChub

slbit

slbit is a zero-dependency Rust crate for luminous semantic packets: portable, deterministic explanation artifacts that sit beside proofs, zkML outputs, agent traces, Rootprint graphs, attestations, and audit logs.

It does not replace proof systems. It answers the question proof systems leave open: what happened, why did it matter, and how should a human or another system inspect it?

Trust Boundary

slbit packets are observation and meaning data. They MUST NOT be used as proof validity, proof soundness, proof identity, Rootprint lineage, replay fingerprint, or cryptographic equivalence inputs.

Changing an slbit packet changes the slbit packet digest. It must not change the identity of the external proof, .pha artifact, Rootprint graph, zk proof, attestation, model output, or agent action it describes.

What 3.1.0 Adds

  • deterministic v3 truth-boundary inspection reports;
  • authority counts, unbound node listing, warning/failure node listing, and transcript node coverage;
  • deterministic semantic dependency-chain and shortest-path inspection;
  • richer local ask questions for proof boundary, branch comparison, dependency review, mutation/failure nodes, and shortest explanation paths;
  • builder helpers for claim cards, graph views, diff views, and proof-status labels.

What 3.0.0 Added

  • slbit/viz-packet/v3;
  • Meaning Observatory packets for verified memory;
  • explicit Memory Capsule, branch, and replay bindings;
  • typed semantic DAG nodes with authority badges;
  • local deterministic ask engine;
  • structured support IDs for answers;
  • non-authoritative LLM context export;
  • digest rejection for semantic mutation;
  • cycle rejection for semantic DAGs.

What 0.2.0 Added

  • slbit/viz-packet/v2
  • stable packet IDs;
  • producer metadata;
  • external anchors such as Power House Rootprint, zk proofs, OpenTelemetry traces, SLSA provenance, model cards, and human approvals;
  • directed acyclic semantic graphs;
  • audience-aware summaries;
  • deterministic privacy redaction records;
  • external signature references;
  • deterministic packet, transcript, and semantic graph digests;
  • fluent Rust builder API;
  • Markdown audit export;
  • v1 verification compatibility.

The crate still has no normal, build, or development dependencies.

Example

use slbit::LuminousPacket;

let packet = LuminousPacket::builder("drone-camera-frame-7842", 4096)
    .producer("drone-perception-demo", "3.1.0")
    .layer("perception-conv3")
    .icon("camera")
    .rgb(0, 200, 255)
    .round(0, &[0xde, 0xad, 0xbe, 0xef], "sensor-processing", "Raw frame converted into features")
    .round(1, &[0x42], "attention-head-7", "Stop-sign feature strongly activated")
    .labeled_node("frame-7842", "input", "Camera frame 7842")
    .labeled_node("conv3", "model-layer", "perception-conv3")
    .labeled_node("attention-head-7", "attention", "attention-head-7")
    .labeled_node("stop-sign-detected", "decision", "stop-sign-detected")
    .edge("frame-7842", "conv3", "processed-by")
    .edge("conv3", "attention-head-7", "activated")
    .edge("attention-head-7", "stop-sign-detected", "supports")
    .anchor_rootprint(
        "drone-perception-7842",
        "rootprint-branch-7842",
        "sha256:aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
        "sha256:bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
    )
    .summary_for("operator", "The drone detected a stop sign and triggered the stop policy.")
    .build()?;

packet.verify()?;
println!("{}", packet.to_markdown("operator"));
# Ok::<(), slbit::SlbitError>(())

Meaning Observatory v3:

use slbit::{BoundCore, MeaningNode, MeaningPacket};

let core = BoundCore::new(
    "phm_earth-001",
    "sha256:aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
    "sha256:bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
);

let packet = MeaningPacket::builder("claim_earth_001", "EARTH-001", "proof-memory", core)
    .node(MeaningNode::new(
        "node_core_001",
        "artifact",
        "Verified core artifact",
        "Power House core verification completed before semantic rendering.",
    ).authority("core"))
    .build()?;

packet.verify()?;
let report = packet.inspect()?;
assert!(!report.semantic_changes_affect_core);
let answer = packet.ask("what is core truth?");
assert!(answer.not_proven_by_this_answer);
# Ok::<(), slbit::SlbitError>(())

v1 Compatibility

The v1 API remains available:

  • LuminousClaim
  • BitInteractiveTranscript
  • SimpleLuminousSumcheck
  • VizPacket

slbit 3.1.0 emits corrected deterministic v1 JSON and still accepts legacy v0.1.0 packet digests during verification.

Documentation

v2 Reference Examples

  • drone_perception_v2
  • zkml_classification_v2
  • financial_audit_v2
  • agent_trace_v2
  • multi_agent_governance_v2
  • medical_ai_decision_v2
  • smart_contract_proof_v2
  • ci_formal_verification_v2
  • supply_chain_provenance_v2
  • human_approval_annotation_v2

License

AGPL-3.0-only.