research/examples/verify_output_provenance.py

#!/usr/bin/env python3
"""Claim 2: output provenance binding.

Bind an output to the certificates of its composition (weight-source and
context-source axes), verify the binding VALID while the composition is
intact, and show it flips to INVALID when a referenced certificate is
revoked or crypto-erased. Also show a tampered output fails the digest check.
"""

import sys
from pathlib import Path

sys.path.insert(0, str(Path(__file__).resolve().parents[1] / "src"))

from capability_licensing import (
    AXIS_CONTEXT_SOURCE,
    AXIS_WEIGHT_SOURCE,
    CertificateAuthority,
    RevocationList,
    RUNG_CRYPTO_ERASURE,
    bind_output,
    verify_output_provenance,
)

failures = []


def check(label: str, condition: bool, evidence: str) -> None:
    marker = "ok " if condition else "FAIL"
    print(f"  [{marker}] {label}")
    print(f"         {evidence}")
    if not condition:
        failures.append(label)


def main() -> int:
    print("=== Claim 2: output provenance bound to composition certificates ===\n")

    authority = CertificateAuthority()
    root = authority.create_root("Reference Root Authority")
    organisation = authority.issue_organisation(root.serial, "Reference Research Organisation")

    weight_source, weight_key = authority.issue_capability(
        organisation.serial,
        "summarisation-adapter (weight-source)",
        claims={"axis": AXIS_WEIGHT_SOURCE, "grant": "research demonstration only"},
    )
    context_source, _context_key = authority.issue_capability(
        organisation.serial,
        "licensed-reference-corpus (context-source)",
        claims={"axis": AXIS_CONTEXT_SOURCE, "grant": "research demonstration only"},
    )

    print(f"  weight-source certificate:  {weight_source.serial}")
    print(f"  context-source certificate: {context_source.serial}")

    output = (
        b"Stand-in generated output: a three-paragraph summary of the supplied "
        b"reference material, produced by the licensed summarisation adapter."
    )
    record = bind_output(
        output,
        sources=[
            (AXIS_WEIGHT_SOURCE, weight_source),
            (AXIS_CONTEXT_SOURCE, context_source),
        ],
        signer_certificate=weight_source,
        signer_private_key=weight_key,
    )
    print(f"  output bound: sha256 {record.output_sha256[:16]}..., "
          f"signed by {record.signer_serial}\n")

    registry = authority.registry()
    revocations = RevocationList()

    # 1. Intact composition verifies VALID.
    intact = verify_output_provenance(
        output, record, registry, [organisation], [root], revocations
    )
    print("  checks performed on the intact composition:")
    for item in intact.checks:
        print(f"    - {item}")
    check(
        "intact composition verifies VALID",
        intact.valid,
        f"result: {'VALID' if intact.valid else 'INVALID'} - {intact.reason}",
    )

    # 2. Tampered output fails the digest binding.
    tampered_output = output + b" [edited after the fact]"
    tampered = verify_output_provenance(
        tampered_output, record, registry, [organisation], [root], revocations
    )
    check(
        "tampered output fails the digest binding",
        not tampered.valid and "does not match the bound digest" in tampered.reason,
        f"result: INVALID - {tampered.reason}",
    )

    # 3. Revoke the context-source certificate: the SAME record now verifies
    #    INVALID. Provenance is a live claim against authority state.
    revocations.revoke(context_source.serial, "context licence terminated")
    revoked = verify_output_provenance(
        output, record, registry, [organisation], [root], revocations
    )
    check(
        "after revoking the context-source certificate, verification is INVALID",
        not revoked.valid and "revoked" in revoked.reason,
        f"result: INVALID - {revoked.reason}",
    )
    revocations.reinstate(context_source.serial)

    # 4. Crypto-erase the weight-source capability: INVALID again, and the
    #    ledger records the strongest rung.
    revocations.revoke(
        weight_source.serial,
        "capability unit crypto-erased by the authority",
        rung=RUNG_CRYPTO_ERASURE,
    )
    erased = verify_output_provenance(
        output, record, registry, [organisation], [root], revocations
    )
    check(
        "after crypto-erasure of the weight-source capability, verification is INVALID",
        not erased.valid and "crypto-erasure" in erased.reason,
        f"result: INVALID - {erased.reason}",
    )

    print()
    if failures:
        print(f"RESULT: FAIL - {len(failures)} check(s) did not behave as claimed")
        return 1
    print(
        "RESULT: PASS - the binding verifies for an intact composition and is "
        "INVALID for a tampered output or a revoked/erased source certificate"
    )
    return 0


if __name__ == "__main__":
    sys.exit(main())