Non-normative simulation reference
This repository illustrates conceptual patterns for Authority Governance (Layer 3)
as defined in the normative standard:
https://github.com/GautierDorval/interpretive-governance-manifest (pinned releases preferred)

This is NOT production code.
This is NOT a prescriptive implementation guide.
This is NOT an endorsement of any specific technical stack.
This is NOT a safety guarantee.

This repository exists to:

• illustrate normative constraints defined in the manifest (Layer 3)
• demonstrate decision flows and data structures
• provide conceptual integration patterns (simulation-only)

This repository does NOT:

• provide executable orchestration code
• prescribe tool wiring or infrastructure
• include security hardening or production patterns
• guarantee safety, compliance, or fitness for purpose
• provide bypass or exploitation guidance

Conformance to the standard is determined by the manifest, not by this reference.

All code under stubs/ raises NotImplementedError on purpose.

This is a safety and scope boundary:

• to prevent accidental production use,
• to avoid implying a normative implementation recipe,
• to keep this repository focused on flows + schemas + examples.

Normative definitions live in the manifest repository:

• Manifest repo: https://github.com/GautierDorval/interpretive-governance-manifest
• Preferred: pinned releases (this demo references v1.4.1 in DEMO-META.yaml).

If a discrepancy exists, the normative repository prevails.

This repository illustrates Authority Governance (Layer 3).
If you also need a minimal executable illustration for Constraintive Governance (Layer 2), see:

• https://github.com/GautierDorval/interpretive-agentic-reference

These two references are complementary:

• Layer 2 reference: bounded retrieval + fixed inference + schema validation + policy-driven abstention
• This repo (Layer 3): authority classification + policy decision + ledger entry (simulation-only)

• authority_request → classification_result → policy_decision → ledger_entry
• decision paths: allow / deny / escalate / simulate
• simulation-only behavior (no execution)
• conflict handling (planner vs critic)
• ledger entry format examples (traceability + opposability signals)

flowchart LR
  AR[authority_request] --> CR[classification_result] --> PD[policy_decision] --> LE[ledger_entry]

Start with:

• SCOPE.md
• flows/authority-request-flow.md
• schemas/
• examples/

Each scenario provides a complete chain of artifacts:

• examples/authority-requests/
• examples/classification-results/
• examples/policy-decisions/
• examples/ledger-entries/

Apache-2.0. See LICENSE.md.