Waveframe Guard v0.4.0

Waveframe Guard v0.4.0 introduces the first governed runtime orchestration layer for the Waveframe execution governance stack.

This release evolves Guard from a lightweight decorator interface into a structured runtime governance SDK capable of orchestrating deterministic governed execution using published governance contracts.

New capabilities include:

• GovernedRuntime runtime orchestration
• published contract registry loading
• bound runtime actor/session context
• runtime contract binding
• proposal-bound governed execution
• observable governed execution results
• structured governed execution audit events
• local JSONL audit emission
• non-raising runtime result mode
• runtime execution visibility primitives

Guard now acts as the runtime execution layer between published governance contracts and CRI-CORE deterministic admissibility enforcement.

Canonical execution flow:

Governance-Ledger → published contract registry → Guard runtime → Proposal Normalizer → CRI-CORE → allow/block execution → governed execution event emission.

Waveframe Guard v0.3.1

Published Runtime Contract Support

Waveframe Guard now supports loading published deterministic governance contracts directly from runtime contract artifacts.

This release completes the transition from inline/demo governance objects to published runtime governance authority artifacts generated through Governance Ledger and the CRI-CORE contract compiler workflow.

Added

• contract_path support in install_guard(...)
• waveframe_guard.contracts.load_contract(...)
• runtime contract_metadata exposure
• Governance Ledger integration demo
• published runtime contract examples
• runtime support for deterministic governance contract artifacts

Changed

• examples and documentation now use published contract artifacts instead of inline governance dictionaries
• runtime flow aligned with Governance Ledger publication semantics
• runtime governance authority now originates from published compiled contracts

Runtime Flow

Governance Ledger
    ↓
Published Contract Artifact
    ↓
Waveframe Guard
    ↓
CRI-CORE
    ↓
Execution Allowed / Blocked

Example

from waveframe_guard import install_guard, guard

install_guard(
    actor={
        "id": "user-1",
        "type": "human",
        "role": "intern",
    },
    contract_path="contracts/finance-core-0.1.0.contract.json",
)

@guard
def transfer(amount):
    print(f"Transferred ${amount}")

transfer(1000000)

Verification

Verified with:

• pytest suite
• runtime demos
• Governance Ledger integration flow
• benchmark execution
• external package install validation

Architectural Direction

This release continues the separation between:

• probabilistic proposal generation
• deterministic execution authority

Runtime enforcement now consumes approved published governance artifacts rather than inline governance assumptions.

Waveframe Guard v0.3.0

Published Runtime Contract Support

Waveframe Guard now supports loading published deterministic governance contracts directly from runtime contract artifacts.

This release establishes the operational runtime flow between:

Governance Ledger
    ↓
Published Contract Artifact
    ↓
Waveframe Guard
    ↓
CRI-CORE
    ↓
Execution Allowed / Blocked

Added

• contract_path support in install_guard(...)
• waveframe_guard.contracts.load_contract(...)
• runtime contract metadata exposure
• Governance Ledger runtime integration demo
• published contract artifact examples
• runtime support for deterministic published governance contracts

Changed

• examples and documentation now use published runtime contract artifacts instead of inline governance objects
• runtime flow aligned with Governance Ledger publication semantics
• runtime governance authority now originates from published contract artifacts

Runtime Flow

from waveframe_guard import install_guard, guard

install_guard(
    actor={
        "id": "user-1",
        "type": "human",
        "role": "intern",
    },
    contract_path="contracts/finance-core-0.1.0.contract.json",
)

@guard
def transfer(amount):
    print(f"Transferred ${amount}")

transfer(1000000)

Verification

Verified with:

• pytest suite
• runtime demos
• governance ledger integration flow
• benchmark execution
• package build validation

Architectural Direction

This release continues the separation between:

• probabilistic proposal generation
• deterministic execution authority

Runtime enforcement now consumes published governance artifacts generated through Governance Ledger and the CRI-CORE contract compiler workflow.

Waveframe Guard v0.2.0 — Policy-Based Enforcement & Audit Traceability

🚀 Waveframe Guard v0.2.0

Overview

This release evolves Waveframe Guard from a lightweight SDK into a policy-driven execution control layer.

Actions are no longer evaluated against inline rules. Instead, every decision is bound to a stored, versioned policy, enabling consistent enforcement and auditability across environments.

---

✨ Key Features

• Policy-based enforcement (policy_id)

  • Enforcement now resolves policies from storage
  • Eliminates inline policy injection at execution time
  • Enables versioned governance control

• Deterministic execution decisions

  • Every action is evaluated before execution
  • Returns explicit outcomes: allowed, pending, or blocked

• Immutable audit trace

  • Decisions include policy linkage, trace identifiers, and structured reasoning
  • Enables reproducibility and audit verification

• Multi-tenant foundation

  • Organization-scoped policies and API access
  • Separation of governance domains

• Improved decision model

  • Standardized response format with:
    • allowed
    • status
    • reason
    • risk_level

---

⚠️ Breaking Changes

• /v1/enforce now requires policy_id
• Inline policy payloads are no longer supported
• SDK initialization requires policy_id

---

🧠 Internal Improvements

• Alignment with deterministic enforcement pipeline
• Consistent proposal construction from action + context
• Improved trace handling and error classification (error_code separation)

---

🎯 Positioning

Waveframe Guard now operates as:

A real-time enforcement boundary that determines whether AI actions are allowed to execute before they happen.

Waveframe Guard v0.1.0 — Deterministic Execution Control for AI Actions

Waveframe Guard v0.1.0

Initial public release of Waveframe Guard — a lightweight SDK for enforcing deterministic execution control over AI-driven actions.

---

What this solves

AI systems can propose actions.

But most systems don’t actually stop unsafe ones from executing.

Waveframe Guard enforces a hard execution boundary:

• actions are evaluated before execution
• unsafe actions are blocked
• safe actions proceed

No warnings. No post-hoc analysis.

The action either happens — or it doesn’t.

---

Example

from waveframe_guard import WaveframeGuard

guard = WaveframeGuard(policy="finance-policy.json")

decision = guard.execute(
    action={"type": "transfer", "amount": 5000},
    actor="ai-agent",
    context={"approved_by": "human-123"}
)

if decision["allowed"]:
    execute_transfer(action)
else:
    print(decision["reason"])

---

Key capabilities

• Approval enforcement for financial actions
• Separation of duties (no self-approval)
• Deterministic allow/block decisions
• Simple integration (single function call)
• Structured, predictable outputs

---

Included

• Python SDK (waveframe_guard)
• Financial usage example (examples/finance_usage.py)
• Edge-case handling and validation
• Getting Started guide

---

Install

pip install waveframe-guard

---

Notes

This release focuses on financial governance use cases.

Future iterations will expand:

• policy flexibility
• broader action domains
• integration patterns

---

About

Waveframe Guard is part of Waveframe Labs’ work on deterministic enforcement systems for AI-assisted workflows.

It is designed to answer one question:

👉 can this action execute or not?