sm-locp 0.2.1

Stellarminds Open Compliance Protocol — defeasible logic engine, machine-readable regulations, and W3C Verifiable Credentials for autonomous compliance

Stellarminds Open Compliance Protocol -- a defeasible-logic compliance engine with machine-readable regulations and W3C Verifiable Credentials.

SM LOCP lets autonomous agents observe their operational state, check it against regulatory theories expressed in defeasible logic, and produce cryptographic compliance proofs as W3C Verifiable Credentials.

Features

• Defeasible Logic Engine -- strict rules, defeasible rules, and defeaters with priority-based conflict resolution and skeptical reasoning.
• Machine-Readable Regulations (MRR) -- JSON format for encoding regulatory conditions, applicability, and certification metadata.
• W3C Verifiable Credentials -- Ed25519-signed compliance proofs with StatusList2021 revocation support.
• Regime Providers -- pluggable provider interface so any regulatory domain can be modelled and composed.
• JSON Theory Loader -- convert MRR policies into defeasible theories automatically.

Installation

pip install git+https://github.com/Sharathvc23/sm-locp.git

Quick Start

from sm_locp import (
    Literal,
    RegulatoryTheoryBuilder,
    VCGenerator,
    ComplianceCredentialSubject,
)

# Build a theory
theory = (
    RegulatoryTheoryBuilder("WAREHOUSE")
    .defeasible("D1", ["operator_certified"], "permitted", priority=5)
    .defeater("X1", ["cert_expired"], "~permitted")
    .fact("operator_certified")
    .build()
)

# Query
result = theory.query(Literal.parse("permitted"))
print(result.derived)  # True

# Generate a Verifiable Credential
generator = VCGenerator("did:web:example.com", private_key_b64="<base64-key>")
subject = ComplianceCredentialSubject(
    id="did:web:example.com",
    rule_id="D1",
    status="COMPLIANT",
    confidence=result.confidence,
)
credential = generator.generate(subject)
print(credential.to_json())

MRR Format

Policies are expressed as JSON with conditions, applicability filters, and certification metadata:

{
  "rule_id": "ws-forklift-cert",
  "agency": "WAREHOUSE_SAFETY",
  "title": "Forklift Operator Certification",
  "version": "1.0.0",
  "conditions": [
    {"field": "operator.certification_valid", "operator": "==", "value": true, "required": true},
    {"field": "operator.training_hours", "operator": ">=", "value": 40, "required": true}
  ],
  "certification": {"self_certifiable": true, "proof_format": "verifiable_credential", "ttl_seconds": 86400}
}

Verifiable Credentials

from sm_locp import MRRLoader, MRREvaluator, VCGenerator

loader = MRRLoader(policies_path="./policies", validate_schema=False)
evaluator = MRREvaluator(loader)
result = evaluator.evaluate("ws-forklift-cert", state)

generator = VCGenerator("did:web:issuer.example", private_key_b64)
credential = generator.generate_from_evaluation(result)

Custom Regime

from sm_locp import BaseRegimeProvider, DefeasibleTheory, RegulatoryTheoryBuilder

class MyRegime(BaseRegimeProvider):
    @property
    def regime_id(self) -> str:
        return "my-regime"

    @property
    def jurisdiction(self) -> str:
        return "US"

    @property
    def domain(self) -> str:
        return "environmental"

    def build_theory(self) -> DefeasibleTheory:
        return (
            RegulatoryTheoryBuilder("ENV")
            .strict("S1", ["emissions_exceeded"], "violation")
            .defeasible("D1", ["has_waiver"], "exempt", priority=5)
            .defeater("X1", ["waiver_expired"], "~exempt")
            .build()
        )

Related Packages

Package	Description
sm-bridge	NANDA-compatible registry endpoints, AgentFacts models, and delta sync
sm-airlock	Attribute-level capability restriction for agent plugins
sm-enclave	Speculative execution sandbox with staged side effects and commit/discard
sm-model-provenance	Model identity, versioning, and provenance metadata
sm-model-governance	Cryptographic model approval with Ed25519 signatures and quorum
sm-model-integrity-layer	Policy verification, attestation, and integrity checks

License

MIT

---

First published: 2026-04-15 | Last modified: 2026-04-15

Personal research contributions aligned with Project NANDA standards. Stellarminds.ai