Maka's Computer Use surface was built against other people's work in three materially different ways. They are separated here because the obligations differ: redistribution under a license, adaptation or reference use of licensed source, and inspection of a proprietary implementation for compatibility.
At a high level, Maka's Computer Use implementation was informed by the
MIT-licensed trycua/cua and iFurySt/open-codex-computer-use projects and by
comparison with Codex Desktop's implementation and externally visible
behavior. Maka integrated and adjusted those inputs through its own testing,
safety boundaries, and product requirements.
That summary does not collapse the source categories below. Licensed source, behavior observations, exact facts recovered from a binary, and Maka-authored work remain separate because their evidence and release-review obligations are different.
Paths are Maka's unless they carry the upstream repository name, as in
open-codex-computer-use/....
Every entry names what was taken, where it landed, and what the evidence was. When you add or change a borrowed design, add the row here and put the same statement in the file that carries it — an in-file comment answers "why is this constant 200?" at the moment someone asks it, and this file answers "what did we build on?" for the project as a whole.
Ships inside the Maka artifact. Requires the license text and copyright notice to travel with it.
| Component | License | Where the notice lives |
|---|---|---|
| npm dependencies | various | apps/desktop/resources/licenses/npm/, generated by scripts/generate-third-party-notices.mjs and byte-checked at build by scripts/check-third-party-notices.mjs |
Maka has one Computer Use executor, maka-cu. It is built from Maka's own
source by the prepare command in scripts/computer-use.mjs and pinned by digest in
apps/desktop/bundled-tools.json. It is not signed, so it is not distributed:
its distributionReady is false and scripts/verify-macos-dmg.mjs
forbids its path in a packaged build. Nothing third-party ships inside the Maka
artifact for Computer Use, which is why the table above has one row.
The third-party executor Maka used to run, cua-driver, was removed along with its vendored notice. Nothing in this tree spawns it or links against it.
maka-cu is itself a fork of MIT-licensed iFurySt/open-codex-computer-use
(§2), so when it does ship, that notice travels with it.
MIT-licensed source was used both as an implementation starting point and as a design reference. The rows below distinguish adaptation from reference-only use.
MIT. The original Maka cursor renderer was a TypeScript adaptation of the
public cursor-overlay crate, introduced in Maka commit
025d0c628a2162d0a7daf49e97d104c36a4431c6. The fixed upstream source recorded
by Maka's bundled cua-driver manifest was commit
8c921b2b3bf13494724ead4f0a814d80c56a7e8b.
Maka later replaced and extended the planner, timing, hotspot, rendering, and presentation lifecycle. The current source remains in that documented lineage even though the old cua-driver executable and its vendored notice no longer ship.
Both MIT, © 2026 Leo. An independent reimplementation of Codex's Computer Use as an MCP server.
For the rows below, no source code was copied into this repository; what was taken is a format, a decision, or an archived measurement. Attribution is given because it was load-bearing, not because MIT compels it for ideas.
| Taken | Landed in | Notes |
|---|---|---|
An archived capture of Codex's real get_app_state result |
packages/runtime/src/computer-use-tools.ts |
Their open-codex-computer-use/artifacts/tool-comparisons/20260417-focus-behavior/. This is what turned Maka's model of Codex's observation format from inference into an observed sample. |
| The one-line-per-element observation shape: indentation for containment, states written only when not the default | same file | Maka's version keeps observation_id in the header (frame binding is protocol here, prose there) and keeps element geometry (Codex has no coordinate action surface to need it; Maka's is disabled by default rather than absent). |
Writing only the informative half of a state — disabled, never enabled |
same file | Their summarizeTraits. |
Filtering AXPress out of an element's advertised action list, because pressing is what click does |
not yet landed — waits on trycua/cua#2622 exposing per-element AX actions |
Their meaningfulActions. |
| The permission-onboarding pattern: anchor a guidance panel to the System Settings window, track it, and distinguish grants that need an app relaunch | not yet landed — for the feat/permission-onboarding work |
Their open-codex-computer-use/apps/OpenComputerUse/Sources/OpenComputerUse/PermissionOnboardingApp.swift. |
| Archiving side-by-side tool captures in-repo as evidence | practice, not code | Worth adopting for Maka's own Codex comparisons. |
| Private SkyLight synthetic-focus/background click recipe | Maka's maka-cu executor, pinned in apps/desktop/bundled-tools.json |
Derived independently from the same MIT cua-driver/yabai recipe; the executor keeps its attribution and notice. The WebContent path uses the host window and a single private post rather than the general dual-post compatibility route. |
Removing the cua-driver binary from this tree changes nothing about that path: what was borrowed was a published recipe, not the artifact or proprietary Codex implementation.
The signed Codex Desktop Computer Use executable was statically inspected. No OpenAI source code or executable is included or redistributed by Maka, but specific geometry, numeric constants, and control-flow facts recovered from the binary were transcribed into the cursor implementation. This is a claim about the listed facts, not a claim that the Computer Use implementation as a whole was derived from the binary.
This inspection provides evidence, not a license grant. Pull requests #1255 and
#1883 accurately recorded most of the listed binary-recovered inputs, but #1883's
"term-for-term" phrase was broader than the current mixed scorer. The detailed
artifact, retained facts, and Maka-authored divergences are recorded in
docs/computer-use-cursor-provenance.md.
| Binary-recovered or observed fact | Informed area |
|---|---|
| Exact cursor geometry, center hotspot, motion configuration, close-enough thresholds, path measurement, and core scoring weights | apps/desktop/src/renderer/computer-use-overlay/engine/cursor-engine.ts |
| Overlay level policy — an occluded target raises the cursor rather than hiding it | same file, and apps/desktop/src/main/computer-use/cursor-overlay-window.ts |
| The observation text shape | packages/runtime/src/computer-use-tools.ts, corroborated by the archived capture in §2 |
| OOP WebContent targeting, retained-element unique refetch, and renderer-generation fencing | Reimplemented in the pinned maka-cu source; this repository records the exact source commit and binary digest in apps/desktop/bundled-tools.json |
| Stable AX ids across revisions, ordered no-change/remove/insert/update presentation, and full-tree fallback | Reimplemented in the pinned maka-cu source and packages/runtime/src/computer-use-observation-text.ts; the complete current tree remains authoritative and dispatch still uses fresh snapshot tokens |
Where Maka deliberately diverges, the divergence is stated at the point of divergence rather than here, so it is read by whoever is changing that code.