clawscan.json, endor.json, skillspector.json, virustotal.json: 5 bytes each (null)
static-analysis.json: present (4.4.5: engine v2.4.27, 19 findings, same six reason codes as the public 4.4.3)
The only difference from the public 4.4.3 is that clawscan.json is null instead of a 3 KB verdict. The static findings are the same family as 4.4.3 (the package is a security audit tool, so its own detection signatures contain strings such as exec( and eval(atob().
This looks like the partial-SkillSpector-report regression described in the “Retained ClawHub releases stuck after incomplete security checks” thread. I am not asking for a waiver or bypass. Please re-run the required scans for the two retained uploads. I have not re-uploaded duplicates and will not.
Thanks for the detailed report, and particularly for preserving the original uploads rather than submitting duplicate versions.
You’re right that this resembles the issue discussed in the earlier retained-release thread. A regression involving partial SkillSpector reports was addressed in ClawScan #129, although we can’t yet confirm that’s the cause of your specific releases being held.
There is a supported way to request fresh scans of existing versions directly, so let’s try that before escalating.
While authenticated as the publisher (clawhub login), run:
The --update option writes successful scan results back to the selected version. It does not require uploading another release.
Once those complete, check whether the versions have become publicly available. A successful scan alone doesn’t necessarily clear an existing publication hold.
If either request fails, or both scans succeed but the releases remain hidden, please share the relevant sanitized scan results and publication status here. No need to repeat the static-analysis work you’ve already done.
And please continue to preserve the existing uploads. We don’t want to bypass security checks or create unnecessary duplicate releases.