Twinned from https://github.com/cfsh/ninefold/issues/895 by tools/beads/import_github.py's reconcile pass.
Implement T4 from design_docs/005b-affected-probe-selection.md (§3). Independent of T1–T3 (all merged: #876, #882/#883, #889/#892).
Goal: one doc-only PR appending an Amendments entry to design_docs/005-probe-selection-and-build-config.md, per Step 7 of .claude/commands/design.md.
Content, all evidence already written up: 1. The three retracted premises, from #755 verbatim: S1's 1.8× (correctness, not cost); cost gated on trust (the worker-pool falsification table); the 34/34 serial sweep (S10 trigger unfired, piece 5b done by measurement — close #757 with this). 2. The S12 deviation: #847 turned the sweep off; the revert condition is T2 of 005b (now merged — #882, #883) — say so. 3. A pointer: remaining work re-planned in 005b.
Done when: the amendment follows Step 7's entry template, body prose is untouched except additive pointers, and @aedanpope has approved (design_docs/** always waits for him).
None.
No comments.