Engineering work

CTOX first. Engineering proof around it.

Metric Space builds CTOX, a runtime and Business OS interface for long-running agentic technical work. Selected public repositories show the surrounding engineering discipline: agent interfaces, code navigation, evaluation and mathematical tooling.

CTOX
01Runtime
02Business OS MCP
03Agent skill
04Audit and approvals

Flagship

CTOX ecosystem

Runtime / Business OS / Rust

ctox

Native runtime and daemon for long-running technical work, with Business OS connectivity for external agents, status, records, runs, artifacts and approvals.

GitHub

Agent skill / Deployment path

ctox-business-os-deploy-skill

GitHub-hosted skill that turns /ctox into the agent interface for installing, connecting and verifying CTOX Business OS access from coding-agent runtimes.

GitHub

Supporting work

Agent tooling and evaluation

Agentic coding / Rust

grepplus-rs

Standard grep with extra commands for coding agents: who-calls, impact, context and brief. It makes code navigation cheaper and more reliable.

GitHub

Evaluation / Deep research

deep-search-benchmark-site

Harness-agnostic arena for live-web deep research: fixed tasks, multiple harnesses and quality judged after collection.

GitHub

Mathematical computing / C++ and Python

metric

Numerical computing framework for finite metric spaces. It is foundational rather than product-led, but it fits the Deep Data Engineering base layer.

GitHub

In development / AI learning systems

learnordie

Lecture augmentation platform for technical courses. It is in active development and not yet presented as finished product work.

GitHub

Previous work

Octopus and the historical GPT-5 system-prompt repository are not part of the current showcase. They remain useful context, but CTOX is the current flagship story.