Deterministic Governance
Deterministic Governance is the architectural practice of enforcing security policies and operational limits through external, hard-coded software gates rather than probabilistic LLM prompts.
“Deterministic Governance replaces probabilistic text instructions with hard-coded code execution gates that cannot be bypassed by model hallucination.”
Probabilistic AI models can never guarantee 100% adherence to natural language instructions. Deterministic Governance guarantees zero unauthorized actions by running policy gates in standard code.
Deterministic Control Plane Interception
Reverse Citations: Implemented & Audited Across Platform
Richard Ewing’s Research Thesis
Relying on natural language prompt rules (CLAUDE.md / .cursorrules) for security is security through optimism. Real governance requires deterministic proxy interception.
Why This Specification Exists
Developers rely on text instruction files (.cursorrules, CLAUDE.md) which models routinely ignore under context rot.
Adding more rules to the system prompt.
Prompt instructions compete for context space and fail during reasoning pressure.
Created an external TypeScript proxy layer enforcing strict JSON schema allowlists.
What Changes If You Believe This?
Move security checks from system prompts into backend proxy middleware.
Prevent un-sanctioned API calls from scaling compute costs.
Define rigid API parameter boundaries for customer-facing agents.
Log immutable, audit-ready cryptographic execution traces.
Specification Maturity & Ecosystem Spread
Recommended Action by Role
Deploy proxy gates in backend services to intercept tool calls before DB writes.
Exogram Proving Ground
Test deterministic control gates against real-time agent payloads.
Latest Publications & Research Activity
How to Prevent Memory Loss in AI Applications
Giving an AI a bigger memory window is like giving a confused worker a bigger inbox.
Claude Search Fails: Prompting Kills Adoption
Frequently Asked Questions
Q:What is Deterministic Governance?
Enforcing security rules via hard-coded backend software rather than LLM prompts.
Inspectable Evidence Ledger
Classified evidence items supporting, extending, or refining this canonical research specification.
| Evidence Item | Publisher | Evidence Type | Strength | Role | Action |
|---|---|---|---|---|---|
| Security Gates for AI Agents | Built In | Production Telemetry | ★★★★★ | Origin | Inspect ↗ |
Recommended Citation
Ewing, R. (2026). "Deterministic Governance." Richard Ewing Research Canon. Available at: https://www.richardewing.io/concepts/deterministic-governance
@article{ewing_deterministic_governance,
author = {Ewing, Richard},
title = {Deterministic Governance},
journal = {Richard Ewing Research Canon},
year = {2026},
url = {https://www.richardewing.io/concepts/deterministic-governance}
}