Skip to content

meridianverity/acceptance-plane

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

The Acceptance Plane™

Action acceptance before impact.

DOI

Public architecture thesis and hardened public-evaluation proof overlay for the Acceptance Plane™ in agentic AI infrastructure.

Prepared by: Meridian Verity Group
Author: Scott Lee
Canonical thesis version: v1.0.0
Public-eval overlay: v1.0.2-public-eval
Release date: 2026-06-11
Canonical DOI: https://doi.org/10.5281/zenodo.20645907

Canonical Definition

The Acceptance Plane™ is an architectural function that determines whether a specific autonomous AI action should be accepted by a protected system before impact, based on current, scope-bound, verifier-ready evidence at the acceptance boundary.

Short form:

The Acceptance Plane™ decides whether an autonomous AI action becomes a real-world consequence.

Repository Status

This repository contains a public architecture thesis, provenance package, and bounded public-evaluation proof overlay. It is not a formal standard, product specification, legal opinion, compliance certification, patent claim chart, implementation disclosure, endorsement, deployment guarantee, or complete security solution.

What changed in the hardened overlay

The original v1.0.0 package was thesis/provenance oriented. The v1.0.2 public-eval overlay adds runnable verification and release-facing pointer locks:

python scripts/verify_manifest.py
python tools/run_public_eval.py
python tools/verify_transparency_bundle.py
python tools/verify_proof_receipt.py --verify-manifest
python tools/independent_recompute.py
python tools/release_gate.py

Expected headline result:

Acceptance Plane public eval: 64 / 64 PASS
independent recomputation: 13 / 13 PASS
release gate: PASS

Release-facing metadata is locked in:

metadata/github_release_body_v1.0.2-public-eval.md
metadata/public_release_copy_v1.0.2.md
metadata/release_pointer_lock_v1.0.2.md

Fast path

python -m pip install -e . --no-deps
python -m pip install cryptography pytest
make qa-full
python -m pytest -q

cryptography is used for the deterministic public-eval Ed25519 receipt profile. The key is intentionally public-eval only and is not a production trust root.

Contents

  • docs/acceptance-plane-whitepaper.md - canonical public architecture thesis.
  • docs/acceptance-plane-in-one-workflow.md - production deployment workflow explainer.
  • docs/ACCEPTANCE_BOUNDARY_SPEC_v1.md - runnable public-eval decision profile.
  • docs/PROOF_CHAIN.md - proof-chain verification guide.
  • docs/HARDENING_NOTES_v1_0_2.md - hardening and integrity notes.
  • docs/RELEASE_ARTIFACT_VERIFICATION.md - source/archive verification guide.
  • docs/PERMIT_RECEIPT_HALTSEAL_SUPERSET_STRATEGY.md - public-safe relationship note for concrete receipt/gateway mechanisms.
  • docs/STRATEGIC_DISCLOSURE_BOUNDARY.md - public/private disclosure boundary for release hygiene.
  • tools/ - public-eval runner, proof receipt verifier, release verifier, and independent recomputation.
  • acceptance_plane_eval/ - deterministic public-eval reference evaluator.
  • vectors/ - 64 ACCEPT/HOLD/REFUSE scenario vectors.
  • receipts/ - generated public-eval results, transparency bundle, and proof receipt.
  • examples/scenario_cards/ - linter-ready scenario cards.
  • schemas/ - JSON schema profile for scenario cards, evidence, actions, decisions, and receipts.
  • adapters/ - shape-only OPA, Envoy, and Kubernetes adapter sketches.
  • dist/ - original v1.0.0 PDF/DOCX release assets.
  • metadata/ - Zenodo/GitHub metadata and publication copy.
  • provenance/ - file tree, QA report, and SHA256 manifest.

Scenario-card linter

python tools/lint_scenario_card.py examples/scenario_cards/deploy_accept.json
python tools/lint_scenario_card.py examples/scenario_cards/hold_approval_required.json
python tools/lint_scenario_card.py examples/scenario_cards/refuse_revoked.json

The linter checks basic card completeness, verifies policy digest binding, evaluates ACCEPT/HOLD/REFUSE, and prints a deterministic receipt digest.

Canonical Citation

Lee, Scott. Meridian Verity Group. (2026). The Acceptance Plane™: The Missing Trust Layer for Agentic AI Infrastructure (v1.0.0). Zenodo. https://doi.org/10.5281/zenodo.20645907

The v1.0.2 overlay is source/provenance and public-evaluation material for the same canonical thesis. It does not replace the canonical DOI.

Source and Provenance

The GitHub release package is intended as a source/provenance package for public architecture materials and bounded public-evaluation artifacts, not a product software release.

Repository URL:

https://github.com/meridianverity/acceptance-plane

Scope Boundaries

The materials are intended for public discussion, citation, education, review, and provenance.

They are not a formal standard, product specification, API specification, legal opinion, compliance certification, patent claim chart, implementation disclosure, endorsement, deployment guarantee, or complete security solution.

License and Notices

This package uses Creative Commons Attribution 4.0 International (CC BY 4.0) for the public architecture thesis, related documentation, figures, and public-evaluation materials.

See the NOTICE file for trademark, patent, and scope notices.

The license applies only to copyrightable public materials included in this release.

It does not grant patent rights, trademark rights, service mark rights, product implementation rights, certification rights, compliance approvals, endorsements, confidential mechanisms, partner-specific rights, or rights to use Meridian Verity Group marks as source identifiers.

“Acceptance Plane™” and “The Acceptance Plane™” are used as Meridian Verity Group framework identifiers and source-identifying marks.

Short Public Description

The Acceptance Plane™ describes the architectural function for deciding whether a specific autonomous AI action should be accepted by a protected system before impact.

Access is not authority.
Permission is not proof.
Execution is not acceptance.

Not a product implementation. Not a standard. No certification right. No patent license.

About

Public architecture thesis and reference workflow for the Acceptance Plane in agentic AI infrastructure.

Resources

License

Contributing

Security policy

Stars

0 stars

Watchers

0 watching

Forks

Packages

 
 
 

Contributors