Replies: 1 comment
|
Your guess is right. Completeness is scored from the chain's own lifecycle records: the signed session open, heartbeats, session close, and an intent record written before each egress. It catches a record that was cut short or had gaps, and it can't see anything that went around the proxy. That's why its best result is LIMITED and never COMPLETE. The grade does say which one it means. Every Completeness line carries its limit in the output, "mediated evidence cannot prove unmediated egress did not occur", plus a reason code like bounded_closed or heartbeat_gap. Proving nothing went around the recorder is a deployment job (containment or network policy), and the receipts can't do it for you. |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
The evidence scorecard grades Authentic, Untampered, Anchored and Completeness, each with its own stated limit, and the limit given for the boundary is that it cannot speak to whatever happened outside what it mediates.
That is the pair I cannot resolve. If traffic that bypasses the proxy is invisible by construction, I cannot see what observable Completeness is scored against. My guess is continuity of the receipt chain, which would show that nothing was dropped from the record, but not that nothing went around the recorder.
Is that what it measures, and if so does the grade itself say which of the two it means?
All reactions