Krv Labs
Open source · In active development

Quality is the
New Currency

Agents write code at machine speed; we still review it at human speed. Topos reads the structure of agent-generated code — not just whether it runs — and returns an auditable verdict.

WHAT TOPOS DOES

Three lenses on one file.

Topos provides structural code quality metrics that allow AI coding agents to autonomously evaluate, iterate on, and optimize the code they generate.

SimpleAST + CFG
ComposableMDG
SecureCPG
THE VERDICT

Beyond pass/fail.

A Code Quality Medal per file — concrete tiers an agent can target, not a vague “make it better.”

Toggle pillars to see the verdict

SLOP

Passes 0 (or fails to parse)

GOLD

Passes all 3 pillars

SILVER

Passes 2 of 3 pillars

BRONZE

Passes 1 of 3 pillars

SLOP

Passes 0 (or fails to parse)

Set preferences to tell agents which medal — or pillar mix — to target under time and token budgets.

TRY IT

It's open source, and we want you to break it.

Topos is in active development. Run it in your editor or wire it into your agent — then tell us where the verdict is wrong.

VS Code Extension

Get structural verdicts inline, as you review the diff.

Try the VS Code extension

MCP Server

Wire Topos into your agent so it optimizes toward the verdict.

Open

GitHub

Read the engine and run it yourself. BSD-3, open source.

Open
Read the thinking behind Topos — “Quality is the New Currency”