HN Showcase
Back to browse
GitHub Repository

AI agent governance framework with audit trails, safety checks, and replayable decision making for autonomous systems

1 starsPython

ExoArmur – Deterministic governance runtime for autonomous systems

by slucerodev·Mar 7, 2026·1 point·0 comments
Visit ProjectView on HN

AI Analysis

●●SolidBig BrainNiche Gem

Verifiable decision replay for autonomous systems, but execution complexity limits adoption beyond safety-critical domains.

Strengths
  • •Deterministic replay mechanism enables independent verification of why actions were approved or denied, solving real auditability gap.
  • •Separates governance policy from decision intelligence—multiple upstream systems can feed the same execution layer.
  • •V2 feature-gated behind flags allows safe incremental rollout of federation and control plane capabilities.
Weaknesses
  • •Heavily depends on orchestration layer upstream; unclear how it integrates with existing agent frameworks (AutoGen, LangGraph, Crew).
  • •Python-only, no SDKs for other languages yet; severely limits enterprise adoption in polyglot shops.
Target Audience

AI/autonomous systems engineers, security teams building agent frameworks

Similar To

OpenAI's governance work (early research) · Anthropic's Constitutional AI frameworks

Post Description

I built ExoArmur because autonomous systems are becoming capable of taking real actions (deployments, infra changes, trading, etc.), but most systems still lack deterministic governance boundaries.

The core idea is that every action must pass through a verifiable execution boundary that produces a replayable evidence bundle.

Pipeline:

Gateway → ActionIntent → PolicyDecisionPoint → SafetyGate → Approval Workflow → ExecutorPlugin → ExecutionTrace → ExecutionProofBundle

This allows you to deterministically replay a decision and verify exactly why an action was allowed or denied.

The repo includes a restrained-autonomy demo where an operator vetoes an action and the system produces a deterministic audit stream.

I'm particularly interested in feedback from people working on agent infrastructure, security, and safety systems.

Similar Projects

Developer Tools●●●Banger

KarnEvil9, a deterministic AI agent runtime

First public implementation of DeepMind delegation paper. Tested on Zork with governance that blocked 'attack'.

WizardryBig BrainZero to One
oldeucryptoboi
103mo ago
AI/ML●●Solid

Bond – Persistent memory and governance framework for Claude AI

Ambitious architecture: persistent hyperdimensional memory (QAIS), deterministic paragraph retrieval (SLA), and a dual-operator Save Protocol that forces human+AI agreement before writes. The practical touches — a React control panel, an AutoHotkey clipboard bridge/counter, and one-command installer — show someone built this to be used, not just theorized. Platform lock (Windows + AHK) and a README heavy on terminology mean it's exciting for niche adopters but not yet plug-and-play for broader audiences.

Big BrainNiche Gem
J-Dub
114mo ago