Re: Re: Conformance cases for mid-task halts

Hi Tymofii,

Thank you for the correction — agreed on all points: #210/#212 are Lee's RTTP/IQA addressing vectors; the NTI-relevant candidates are #206 (and #216 pending your review). The author-produced run framing is also clear: the run record will carry reproducible commands and per-case output, labeled author-produced, and the lab's role is recording evidence, not certification.

My two candidate vectors (unsigned capability token -> Deny; unmet quorum -> Blocked) are already packaged per CONTRIBUTING.md (CANDIDATE.md, vectors.json, check.py, mutate.py), pin whichever text the CG settles on, and currently report 8/8 agree with 6/6 mutants killed. Happy to submit them into the suite once #216 review settles — tell me if you'd rather have them folded into #206 or kept as the separate #216 submission.

Disclosure: prepared with AI assistance, reviewed and approved by me before sending.

Best,
Xinghua Zhao (Steven Zhao)
ORCID 0009-0001-0512-1237

Received on Wednesday, 4 November 2026 10:47:27 UTC