Stop Claude Code from doing irreversible damage. Policy-gated execution + receipts so you can ship agents without sweating production.

Deterministic agent runtime with explicit plans, typed tools, permissions, and replay. Implements Google DeepMind's Intelligent AI Delegation framework. Named after ELP's Karn Evil 9.

HELM OSS — Open-source core of the HELM Autonomous OS. Policy enforcement, kernel runtime, proof graphs, and audit infrastructure.

Execution OS with human approval gates and pressure-driven task control.

Offline-first LLM orchestration CLI for planning and executing human-approved tasks via validated git patches, with reproducible runs and full auditability.

A kernel-userland protocol enforcing information-theoretic bounds on AI adaptivity leakage, benchmark gaming, and capability spillover.

AIΩN – A formal framework for deterministic, history-native computation.

Example skills, manifests, and reference projects for building on Inactu.

Secure execution substrate for immutable agent skills with explicit capabilities, cryptographic provenance, and auditable deterministic runs.

Built to answer: How do you execute millions of smart contract transactions — concurrently — without breaking determinism?

Deterministic execution boundary for AI systems enforcing signed approvals, replay protection, and cryptographic receipts.

MCP server for deterministic task execution via Domain Profile Agents + Task Templates. Provides planning/orchestration layer for CLI AI assistants (Warp, etc.)

R-LAM is a reproducibility-constrained execution framework for Large Action Models in scientific workflow automation. It enables adaptive, agent-driven workflow execution while enforcing strict guarantees on auditability, determinism, and replayability.

The Universal Goal Execution Model. A deterministic computational substrate for autonomous agents and post-procedural software. Stop programming execution. Start declaring intent.