Hi Justin,

We have now run an initial provenance and license check on commit 
1518442865763d603571f5d77c13ffc3b0c96125.

The repository’s production dependency notice check passed. ScanCode Toolkit 
32.5.0 scanned 2,542 tracked files locally. SCANOSS 0.40.2 scanned 2,231 
code-like files and reported matches in 417 files. Of those, 392 matched Maka 
itself. We reviewed the other 25 manually.

None of those external matches pointed to GPL, AGPL, commercial, or 
source-available code that needs to be removed. They were documented MIT or 
Apache-licensed material, short conventional code, or low-confidence false 
positives.

We also ran FOSSA CLI 3.17.16 in local static mode. It produced a dependency 
analysis, but we did not use FOSSA’s cloud snippet-scanning service, so that 
result does not tell us anything about snippet origin.

This does not settle authorship. Twenty-nine bundled Markdown Skills still lack 
individual source records, and SCANOSS does not normally fingerprint those 
files. We are tracking them in 
https://github.com/maka-agent/maka-agent/issues/2669.

Our current understanding is that Raft orchestrated Claude and Codex during the 
initial bootstrap. We are confirming that with the initial contributor and 
checking the applicable output terms.

There is also a separate Computer Use question. Existing provenance records say 
that some constants were recovered from a proprietary binary. We have asked the 
contributors whether those statements are accurate and where the relevant 
implementation details came from.

The scan is tied to the commit above. We will run the relevant checks again 
against the code-transfer or release-candidate commit. We are also preparing a 
sanitized report with the commands, versions, scope, exclusions, findings, 
manual review, and limitations.

If we cannot establish the source and licensing of any material, we will remove 
it or recreate it independently before the first Apache release. We will keep 
the remaining work public and report its status in the first podling report.

Best,
Yuhan Lei
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to