The rule now has
a visible surface.

UTF-8 Markdown source rendered with mojibake when the static response did not declare a charset.

Routed all eleven adoption Markdown artifacts through an application response that explicitly declares text/markdown; charset=utf-8 and nosniff.

On 2026-07-25, the production one-page-summary path returned HTTP 200 with Content-Type: text/markdown; charset=utf-8 and X-Content-Type-Options: nosniff. The response body was also checked against the mojibake signature and the corrected review language. Inspect the referenced receipt ↗

The defect affected presentation, not the underlying document text. Other non-Markdown downloads were not implicated.

The phrase “nine independent AI evaluations” converted creator-directed cross-model review into a claim of independence and validation.

Replaced the claim with “nine cross-model AI review sessions” and named the boundary: creator-directed, model-assisted review is not independent validation.

The one-page summary, hostile FAQ, contributor record, and civic-accountability mapping were reconciled; a regression test rejects the prior phrase.

Convergence across models remains developmental evidence. It does not substitute for an outside human reviewer, field study, audit, or certification.

The sites described how corrections should work but provided no visible current corrections ledger.

Added this routed ledger, placed it in primary navigation, connected it to Archive and Provenance, and supplied a durable reporting path.

The route and its named records are included in rendering and navigation tests. The ledger itself is the public surface that was missing.

A visible ledger improves traceability; it does not prove every defect will be found or every disputed claim will be resolved in the reporter’s favor.

Public surfaces alternated between “Phillip Linstrum” and “Charles Phillip Linstrum” without explaining the relationship.

Adopted Phillip Linstrum as the public display byline. Formal and historical source records may retain Charles Phillip Linstrum, with an explicit same-author note where the two surfaces meet.

Current routed footers and adoption-facing identity language are reviewed against the convention; preserved historical artifacts are not silently rewritten.

This is identity reconciliation, not a change of authorship or an attempt to erase the original signed record.

The flagship summary said the public repository served v0.7.1 and promised that v1.0 publication was “in progress.”

Replaced the dated promise with a dated observed state: the public repository identified v0.9.5 on 2026-07-25, while the adoption package carries a separately labeled v1.0 candidate.

The repository README was checked directly and the summary now distinguishes repository publication from package state.

The repository remains the authority for its current release. This ledger does not promise a future publication date.

Two exploratory DOCX reports and an interactive viewer could be mistaken for a current, reproducible validation package despite incomplete public custody and evaluation receipts.

Withdrew the three files from direct public download into recoverable project custody and replaced them with a corrected summary plus a static single-run preserved trace.

The public source manifest now contains 13 SHA-256 entries. The three withdrawn filenames are absent from the public library and preserved under the site custody record.

The withdrawal narrows the public claim; it does not erase project history or establish that the corrected trace validates PBHP.

The 127-entry district instrument was directly public while its mixed-review boundary did not close the need for entry-level source and interpretation review.

Preserved the full instrument in recoverable project custody and replaced its public path with a containment notice directing readers to the current routed IN-6 desk.

The public /in6.html path resolves to the custody notice; the routed /in6 desk carries the current official-source comparison and the held file remains in the repository custody directory.

Containment is not a finding that every historical entry is wrong. It prevents an uncompleted review from being mistaken for current publication approval.

The six public sites carried different surface dates and no single human route explaining currentness, evidence cutoffs, custody limits, and correction paths.

Synchronized the public release boundary, added the Project Shadow release map and Human Door field brief, updated current evidence through August 2 where stated, and retained Project Shadow's beta status.

Each site carries the August 3 release band linking the release map, Human Door, and this corrections ledger. The national deep archive now renders 4,786 entries and 6,808 citations from 4,779 embedded records plus its preserved correction transform.

Synchronization does not create certification, independent validation, licensing, production trust, outside review, or a global green.

The current national record was thin after July 17, and the Major Events interface had regressed from dated events into one summary control per year.

Added 33 sourced records and 94 direct citations across July 18–August 3, restored individual dated Major Events cards with exact-entry navigation, and retained the rail on small screens.

The embedded archive now contains 4,779 records and 6,791 citations; the preserved correction transform renders 4,786 records, 6,808 citations, and 379 Major Events. Automated gates require coverage on every date in the currentization window and reject the year-summary control.

Daily coverage in this 17-day window is not a claim that every consequential event was captured, that every earlier calendar day is complete, or that model review constitutes independent validation. Corrections and source challenges remain open.

The August 3 public surface had become stale. The IN-6 desk also retained an obsolete general-candidate-file caveat and overstated the legislative stage of 16 community-project requests.

Appended four bounded national records with 15 direct citations, replaced the current brief, added three sourced Indiana/local implementation records, used Indiana's August 12 candidate list, and rebuilt the 17-project ledger from the bills' current legislative positions.

The national source layer now contains 4,783 records and 6,806 citations; the preserved correction transform renders 4,790 records and 6,823 citations. The IN-6 desk names Wirth and Shreve from the updated state file and distinguishes 15 House-calendar projects, one House-passed/Senate-received project, and one Senate-calendar project, with zero enacted, obligated, or disbursed.

The four national additions do not establish daily completeness. Candidate-list presence does not predict election outcome. Legislative position does not prove funding or delivery, and local implementation remains open.

The August 14 desk no longer reflected the negotiating, military, Justice Department, Gaza, vaccine, and Indiana flood states visible by August 16. The supplied commentary transcript also mixed source leads with unauthenticated imagery, anonymous allegations, insults, and unsupported appearance or medical inferences.

Used the transcript only as a discovery lead; independently checked the strongest claims against AP, Reuters, CNN, CENTCOM, NWS, WFYI, official state records, and bill-status feeds. Appended five bounded national records with 16 citations, replaced the four-record front desk, expanded the local emergency record statewide, and excluded claims that did not clear the evidence boundary.

The national source layer now contains 4,788 records and 6,822 citations; the preserved correction transform renders 4,795 records and 6,839 citations. The prior 4,783-record prefix remains byte-identical. IN-6 flood conditions are bounded to the noon NWS warning and 11:35 a.m. WFYI log; six FY27 bill vehicles and H.R.7008 were rechecked without changing their underlying action dates.

This is bounded consequential-event coverage, not a complete live-news feed. The aircraft-flap image, USS Boxer shortage claim and denial, appearance and medical speculation, and specific Odesa video were not independently authenticated and are not published as facts. Emergency conditions and all time-sensitive claims require later re-verification.

The daytime August 16 edition predated Trump's announced instruction to substantially reduce U.S. participation in Ulchi Freedom Shield. The supplied transcript correctly surfaced the instruction but overstated it as an immediate stop and destruction of the alliance, mixed Trump's Iran assertion with established fact, and added unsupported business-causation, nuclear-history, force-size, and mental-state claims.

Used the transcript as a discovery lead; checked the core event against AP, Reuters, and the Congressional Research Service; appended one bounded national record with three citations; promoted it to the five-record Today desk; and issued the separately identified TR-W26-05A evening supplement without rewriting the earlier five-record currentization receipt.

The national source layer now contains 4,789 records and 6,825 citations; the preserved correction transform renders 4,796 records and 6,842 citations. The prior 4,788-record daytime layer remains an exact prefix. The published record distinguishes the verified reduction instruction from unknown implementation and from formal treaty termination, troop withdrawal, or alliance dissolution.

The exercise began as scheduled, but the practical U.S. reduction and South Korea's fuller response remained pending. The transcript's alliance-destruction headline, immediate-stop claim, quid-pro-quo inference, false account of North Korea's first nuclear test, unsupported troop total, and appearance or mental-state commentary are not published as facts.

The preserved August 14 R1 public outer contained the exact nested Generic Myth v0.1.1 payload whose own metadata classified it PRIVATE_INTERNAL, EXCLUDED_FROM_PUBLIC, and ineligible for public manufacture. The outer release verifier established custody but did not recursively enforce that nested publication classification.

Preserved the August 14 bytes and public history, marked that release superseded only for a packaging-boundary correction, built Project Shadow R1.0.1 deterministically with all 27 operational descendants byte-identical and no Myth payload embedded, and moved Generic Myth v0.2.0 and Full-Canon Myth v0.3.5 to separate optional default-off companions.

The corrected outer is 5,731,663 bytes with SHA-256 6f6f1e16d5e9a20e62403f14af7ce8629ce2d702528fb7f80aaf4a14deb7a1d1. Its admitted Myth-free inner is 5,463,189 bytes with SHA-256 c8c32b12432c954b1a6f852c0c9f81bbbd40167e936be057d4c3de1a0aa3a623. Recursive inspection, exact prohibited-hash denial, deterministic double builds, verifier mutation tests, runtime replay, and Generic v0.2.0's 32/32 tests passed. Anonymous GitHub and Hugging Face ZIP and checksum redownloads matched exactly; both downloaded R1.0.1 archives and Generic companions passed their bounded verifiers; and corrected-boundary checks passed across all six GPT Sites. The retained redownload and six-site receipts establish packaging-correction effectiveness only. Inspect the referenced receipt ↗

Implementation proves a packaging correction and bounded control behavior only. It does not establish efficacy, safety, certification, legal compliance, independent validation, production readiness, or operational deployment authority.