TrackSDOA.US — SDOA in the wild

Software That Organizes Itself

Your AI writes code faster than you can review it. SDOA makes it write modules instead — typed, sandboxed, self-describing, and safe to run. Then it tracks them in the wild.

SDOA, tracked in the wild

Real modules recognized across real codebases.

loading…
Modules tracked
Projects scanned
Module types
SDOA-Verified repos

Modules by type

Recent activity

Run it on your code. No harm, no foul.

The SDOA scanner never overwrites anything. It analyzes your repo and shows what's reusable, what's duplicated, and what could be a module. If it surfaces something useful — great. If not, you lose nothing.

01 · SCAN
Point the scanner at any repo. Read-only. It walks JS/TS/Python/Rust and finds your modules.
02 · DIAGNOSE
Get a report: recognized modules, reuse candidates, harvest opportunities, by type.
03 · ADOPT
Keep what helps, ignore the rest. Earn the SDOA-Verified badge when your manifests check out.
✓ SDOA-Verified

Built for humans and their agents

The difference between "AI wrote some code" and "AI assembled a verifiable system."

🧠 Self-describing modules

Every module carries a manifest: id, type, inputs/outputs, dependencies, lifecycle. Your system explains itself.

🤖 Agent-composable

Agents plan against a typed capability catalog and run on a deterministic engine — composing vetted blocks, not emitting raw spaghetti.

🧩 Sandboxed & sovereign

Strict boundaries and capability flags. A module can't quietly reach the network or the filesystem unless it declares it.

🛠️ Safe evolution

Refactor one module without breaking the world. Versioned, signable, swappable.