AI Coding Agent Memory & Semantic Change Control

The coordination benchmark, adjudicated.

An independent benchmark of Nool against git-based coordination baselines, built so a skeptic can check every number: each claim is stated so it could fail, with the evidence, and with what it does not establish.

Bare Nool: 709.4 vs. 624.4 valid tasks/hour against git worktrees, 77 vs. 441 child processes spawned.
Full governance ties git worktrees on throughput (632.2 vs. 624.4, p = 0.35) instead of costing coordination speed.
Governance is off by default: none of the three built-in recovery checks catch a change that already landed.

Nool: AI Engineering Governance & Agent Reliability

The governance layer for AI-native engineering teams

Nool governs each repository and coordinates the workspace above it. Engineering leaders can turn cross-project goals into real child-ledger tasks, roll release health and operating metrics upward, and trace delivery back to signed implementation evidence.

Built for engineering managers, CTOs, and principal engineers who need one recursive control model from organizational intent to project execution without creating a shadow management ledger.

Key capabilities: workspace topology reconciliation, goal-to-task decomposition, conservative release-health rollups, child-ledger insights, dependency-ordered coordination, semantic blast radius analysis, provenance, and policy-gated changes.