[pinterbanget]

Summary

This is a portable-replay repair of packages authored from this account (pinterbanget, Ryanson Jonathan). It follows the maintainers' #4470 review (docs/research/pr-intake-evidence/pr4470/PROOF-REVIEW.md), which found two problems: my #4466 reads two provenance files absent from its VERIFY.json inputs, and my #4478 exceeded its 120 s timeout on first run. An audit of all 59 packages from this account found the same pattern elsewhere, plus two other replay breaks on current main.

No scientific claim, number or conclusion changes. No new κ.

Fixes

1. Missing VERIFY.json inputs, 18 packages. These verifiers hash every SOURCES.json entry, including the provenance files certificates/selected-result.json and docs/research/agent-task-catalog.json, but did not list them as inputs. Exported replays therefore failed. The files are now listed.
2. Disposition receipts that failed on current main, 3 packages. halo-chain-receipt (Disposition receipt for the Gaussian halo-exponent PR chain (13 open PRs; none retained; one unchecked) #4473), p17-chain-receipt (Disposition receipt for the p17 κ ≈ 0.003 PR chain (stacked content, pin-only differences, docs conflict with #4436) #4474) and direct-flow-chain-receipt (Disposition receipt for the #4436 direct-flow chain (6 PRs; one withdrawn claim recorded) #4477) asserted that the listed packages were absent from the working tree. Those PRs have since merged. Retention is a fact about the recorded base commit, so the verifiers now check the recorded at-base data instead.
3. Working-source drift on the selected record, 4 packages.
   • pr4100-bit-moment-enclosure, pr4270-outer-exponent-check and pr4338-b2-chain-check read the selected record as of their base: blob b1fea0c9, before main reselected κ = 0.0023. That historical record is now vendored as SELECTED-AT-BASE.json, SHA256-pinned in code and marked -text, instead of being pinned against the live file.
   • pr4326-bit-root-check listed the selected record as an input but never read it, so that unused pin is removed.
4. Timeout. pr4470-arithmetic-check (Independently check #4470's arithmetic (κ = 0.00302015): invoice, local change, root bracket, tails (no new κ) #4478) goes from 120 s to 300 s; it measured 43 s locally.

Evidence

• All 59 packages from this account pass --check-only; before this change, one did not.
• The input audit, comparing each package's hashed SOURCES.json entries with its VERIFY.json inputs, now finds 0 omissions; before it found 18.
• The registered replay of all 21 changed packages passes. The log is in the PR comment.

Head: 02920a630738c5b89f615cf859b40020f02c498f. Base: canonical main f626db5c.

Prepared by Ryanson Jonathan with substantial Claude Code (Claude Opus 5.5) assistance.

🤖 Generated with Claude Code

[pinterbanget]

Registered replay of all 21 changed packages at head 02920a63 (local, Windows, Python 3.12):

  research/column-cut-check 0 5 PASS_REGISTERED_RESEARCH_PACKAGE research/column-cut-check outcome=completed
  research/cross-rounding-check 0 1 PASS_REGISTERED_RESEARCH_PACKAGE research/cross-rounding-check outcome=completed
  research/delta-cross-cell-transfer 0 43 PASS_REGISTERED_RESEARCH_PACKAGE research/delta-cross-cell-transfer outcome=completed
  research/delta-halo-exponent-check 0 21 PASS_REGISTERED_RESEARCH_PACKAGE research/delta-halo-exponent-check outcome=completed
  research/direct-flow-chain-receipt 0 1 PASS_REGISTERED_RESEARCH_PACKAGE research/direct-flow-chain-receipt outcome=completed
  research/eps-halo-exponent-check 0 1 PASS_REGISTERED_RESEARCH_PACKAGE research/eps-halo-exponent-check outcome=completed
  research/gauge-column-check 0 1 PASS_REGISTERED_RESEARCH_PACKAGE research/gauge-column-check outcome=completed
  research/halo-chain-receipt 0 1 PASS_REGISTERED_RESEARCH_PACKAGE research/halo-chain-receipt outcome=completed
  research/p17-chain-receipt 0 1 PASS_REGISTERED_RESEARCH_PACKAGE research/p17-chain-receipt outcome=completed
  research/pr4100-bit-moment-enclosure 0 1 PASS_REGISTERED_RESEARCH_PACKAGE research/pr4100-bit-moment-enclosure outcome=completed
  research/pr4100-chirp-odometer-tariff 0 2 PASS_REGISTERED_RESEARCH_PACKAGE research/pr4100-chirp-odometer-tariff outcome=completed
  research/pr4100-crt-repair-tariff 0 54 PASS_REGISTERED_RESEARCH_PACKAGE research/pr4100-crt-repair-tariff outcome=completed
  research/pr4100-guard-coloring 0 157 PASS_REGISTERED_RESEARCH_PACKAGE research/pr4100-guard-coloring outcome=completed
  research/pr4100-proof-composition-audit 0 9 PASS_REGISTERED_RESEARCH_PACKAGE research/pr4100-proof-composition-audit outcome=completed
  research/pr4100-union-head-travel 0 4 PASS_REGISTERED_RESEARCH_PACKAGE research/pr4100-union-head-travel outcome=completed
  research/pr4270-outer-exponent-check 0 1 PASS_REGISTERED_RESEARCH_PACKAGE research/pr4270-outer-exponent-check outcome=completed
  research/pr4326-bit-root-check 0 1 PASS_REGISTERED_RESEARCH_PACKAGE research/pr4326-bit-root-check outcome=completed
  research/pr4338-b2-chain-check 0 1 PASS_REGISTERED_RESEARCH_PACKAGE research/pr4338-b2-chain-check outcome=completed
  research/pr4470-arithmetic-check 0 44 PASS_REGISTERED_RESEARCH_PACKAGE research/pr4470-arithmetic-check outcome=completed
  research/stage-zero-forced-translation 0 2 PASS_REGISTERED_RESEARCH_PACKAGE research/stage-zero-forced-translation outcome=completed
  research/three-base-census-recount 0 1 PASS_REGISTERED_RESEARCH_PACKAGE research/three-base-census-recount outcome=completed
  DONE