Research consolidation: authenticate all PR99 files retained through PR103 (no new kappa)

[bbMisk]

Change and mathematical scope

Research consolidation; no new kappa. At the research base, PR99 has no integration location in the ledger and Divjot Singh Manchanda has no individual entry in the root contributor list, although incorporated PR103 says it retained the entire contribution.

This package authenticates that claim: all 46 PR99 files have the same complete Git tree in PR103, including paths and modes. An offline proof built from 32 raw Git objects also checks that the original PR99 commit contains exactly those 46 additions and no other changes, and connects PR103 to the published source archive. The contributor entry now links this evidence and preserves separate mechanism and integration credit.

A separate saved-report reconciliation derives the exact profile delta -177100 H1 -31878 H507 -24794 H509 +56672 H511 from the 22 recorded frame extensions, checks zero rank-mass change, and checks that PR103 retains the complete candidate histogram. This is source identity and saved accounting, not a fresh physical replay or new mathematical acceptance. Existing ledger dispositions, frozen inputs and the selected frontier are unchanged.

Task and overlap

OS-random seed 99, consolidation track, overlap-receipt card refined to PR99 through PR103. Task and evidence; coordination claim.

• Research base: bce54fb3f895cbe5cd4c0eb477c3afeed100190f.
• Submission base: 410026dc5513b87db92ae98571d1d5984c69a983.
• PR99: 0622dcce6a897567e4eb6bdfedc7def66ab21346; comparison parent 56b66d58297deca1d7dd130247d720e960f77a37.
• PR103: 75ab1f70d146c21c7611ce0df9eb5f645f51ac0e.
• Retention archive: b91de239a64d24d8aee409a0f4ac3e00710ae5c1.

The final full overlap snapshot covers 785 open PRs, 144 Discussion records and 327 local records with complete remote pagination and no warnings. The only exact-scope matches were this package and its own claim; search limits remain recorded. PR279 source closure (#1190) and current-result routing (#1232/#1285) were rejected as already-covered alternatives. Adam Gall's #1190 uses a related Git authentication method on a different parent package; no code or proof data is imported from it. PR103 and the maintainer reconciliation are credited retention predecessors. Both source PRs were already closed before this work.

Validation

python3.12 scripts/research_package.py --package research/pr99-retention-seed99

Observed on CPython 3.12.13: PASS_PR99_RETENTION_AND_SAVED_ACCOUNTING files=46 singleton_delta=-177100 mass_delta=0, all 18 tests pass (one positive, 17 failure controls), and PASS_REGISTERED_RESEARCH_PACKAGE. Direct optimized checks also passed on CPython 3.14.6.

The controls reject missing/substituted files and modes, missing trees, corrupt source objects, altered archive/review bindings, duplicate ranks/slots, missing charges, and a wrong profile that still preserves total rank. A source export without .git and with PATH=/nonexistent passes. Git's native inventory and full commit diff independently agree; recollecting the proof produces identical bytes and compares all 46 original/retained payloads. Package pins, selected-result verification and whitespace checks pass after rebasing onto the submission base.

No historical full physical/native replay, selected-frontier cold replay or Lean build was run. The default verifier uses no Git, network, third-party library or predecessor implementation. Its source-identity conclusion relies on Git SHA-1 collision resistance; it does not establish signed authorship. Compiler, frame legality, routing, precision/recovery and analytic obligations remain inherited.

Credit

Divjot Singh Manchanda (@djsmanchanda), PR99's nondegenerate saturation; Rohan Arun (@rohanarun), PR103's signed/balanced integration; Zhihao Chen (@jacklightChen), Swapnil Jain and the retained upstream lineage; Douglas Colkitt and the maintainer reconciliation for source preservation. The original NOTICE is preserved byte for byte.

Prepared for Hanson Sun (@bbMisk), with substantial OpenAI Codex assistance in investigation, implementation, verification and writing. No independent human peer review is claimed. Apache-2.0.

Open next questions

No further retention experiment is justified: the complete original change is accounted for. The receipt does not decide whether every original mathematical obligation has received sufficient review; acceptance remains separate. See the package's scope statement and structured task metadata.

[Mister-Meeseeks]

100-PR batch review at 04e7c062789286221fd95938091558b1723015e6.

Merged to main, preserving the original contributor commits and package attribution.

Reviewed scope: Exact authenticated Git-tree and saved-accounting proof that all 46 PR99 additions survive through PR103, preserving individual contributor credit. Source retention and historical histogram equality are distinct from new mathematical admission.

The pinned scoped command passed; the batch ledger contains its command, input hashes, environment and log. This incorporates the research/tooling contribution and does not change the selected κ or discharge the inherited multiplication-theorem assumptions.

Batch review, attribution and evidence. Thank you for the contribution.