moira

Pancha Pakshi Admission Evidence — 2026-07-20

Constitutional Phase 8 Aggregate Intelligence / Eight-Phase Closure — 2026-07-21

The Phase 8 decision has canonical SHA-256 b193b3ba62d1c5eb57d526777310fa16f29c81fa999e69b813670c680ba2fd13. It binds the exact Phase 7 decision at 401c90b7d7c15663427e034a527f983054800f948019cf12b404a0086b3203be, the Phase 6 hardening decision, Phase 4 policy decision, and unchanged manifest.

Phase 8 derives one frozen aggregate from the 24 Phase 7 profiles. It records activity coverage 5/5/5/4/5, all 24 profiles as not evaluable, relation subsets of 17 detected, 7 not recorded, 10 unresolved, 7 named-surface, 0 admitted, and 0 scored, and blocked verse 250. Repeated materialization is identical and the vessel is immutable.

The aggregate rejects incomplete profile, activity, evaluation, or relation partitions; noncanonical activity order; admitted or scored relation drift; and loss of the blocked conflict. Counts remain structural facts only. They are not ranks, weights, favorable/unfavorable judgments, condition scores, prognoses, or medical evidence.

Phase 8 adds no source atom, resolved relation meaning, evaluative doctrine, temporal selection, selector attribution, runtime data, profile, manifest entry, package export, facade method, REST route, prediction, prognosis, diagnosis, advice, or medical-truth claim. The agreed private Phases 1–8 are now complete. There is no automatic Phase 9 transition and public exposure remains unauthorized; either requires a new explicit decision.

Constitutional Phase 7 Integrated Local Condition — 2026-07-21

The Phase 7 decision has canonical SHA-256 401c90b7d7c15663427e034a527f983054800f948019cf12b404a0086b3203be. It binds the exact Phase 6 decision at b175bcd1e537fb551cd26b18d6e6caa37f7a574b7e0a96b336d6fbb97eff9b12, the Phase 5 relation decision, Phase 4 policy decision, and unchanged manifest.

Phase 7 materializes 24 immutable local profiles. Each profile pairs one classification with its exact relation record under the mandatory explicit activity/ordinal policy. The implementation proves identity and source-binding agreement, canonical ordering, complete relation coverage, and one shared policy. Cross-layer substitution, reordered profiles, non-profile members, and injected admitted or scored relation state fail closed.

Every profile is explicitly not_evaluable_no_admitted_condition_doctrine. Favorability is unassigned, condition score is null, prognosis is not performed, medical use is forbidden, and admission remains research-only. Relation detection stays inspectable, but it is never promoted into an evaluation. SLEEP ordinal 5 and verse 250 produce no local profile without fallback or conflict repair.

Phase 7 adds no source atom, resolved relation meaning, generic judgment, numeric score, temporal selection, selector attribution, runtime data, profile, manifest entry, package export, facade method, REST route, prediction, prognosis, diagnosis, advice, or medical-truth claim. Phase 8 may aggregate only stored structural facts unless a separate doctrine is lawfully established.

Constitutional Phase 6 Relation Hardening/Inspectability — 2026-07-21

The Phase 6 decision has canonical SHA-256 b175bcd1e537fb551cd26b18d6e6caa37f7a574b7e0a96b336d6fbb97eff9b12. It binds the exact Phase 5 relation decision at e8e189f75418cc96bc6930e2e93d2cfcebc849cb4080001ee4b4b07b158908d1, the Phase 4 policy decision, and the unchanged manifest.

Phase 6 adds derived inspectability and explicit cross-vessel hardening only. The relation corpus now exposes exact identities, source bindings, presence counts, typed activity filtering, exact activity/ordinal and verse lookup, and separate detected, not-recorded, unresolved, named-surface, admitted, and scored subsets. All views preserve stored Phase 5 ordering and semantics.

The distinction is material: 17 records are detected because their source clauses are present, but 0 relation meanings are admitted and 0 relations are scored. Seven records are explicitly not recorded, ten remain unresolved, and seven retain named bounded surface categories. None of those categories is silently promoted to resolved doctrine. Non-record corpus members now fail with an explicit type error before identity inspection.

SLEEP ordinal 5 and verse 250 return no relation record without fallback or conflict repair. Phase 6 adds no source truth, endpoint, direction, condition, score, temporal selection, selector attribution, runtime data, profile, manifest entry, package export, facade method, REST route, prediction, prognosis, diagnosis, advice, or medical-truth claim. Phase 7 may consider a separately governed integrated local condition, but inherits no condition or scoring doctrine from this closure.

Constitutional Phase 5 Relational Formalization — 2026-07-21

The Phase 5 decision has canonical SHA-256 e8e189f75418cc96bc6930e2e93d2cfcebc849cb4080001ee4b4b07b158908d1. It binds the exact Phase 4 decision at 4a444c91bab9a4949664e6bca4e64ad0ee341b439019db831429e4548bd2c4f9, the Phase 2 closure at a5cd64696d4c040554f2c235056dfd28477fd0796fc82306f44ae43473d434e2, the unchanged manifest, and every Stage 2Q through Stage 2U source decision.

Phase 5 creates one private typed relation record for every unconflicted classification. The 24-record corpus preserves 17 present clauses and 7 not-recorded clauses. The five EAT and five WALK clauses remain unresolved with their source confidence intact. Seven RULE, SLEEP, and DIE clauses retain only their exact bounded surface categories. A surface category does not establish relation endpoints, direction, normalized semantics, condition, or score.

The corpus proves that relation presence is an exact projection of the Phase 2 semantic marker, every record keeps its source decision binding, and canonical activity/ordinal order is unchanged. It rejects not-recorded repair, unresolved normalization, invented confidence, source-binding substitution, and record reordering. Verse 250 remains outside the relation corpus as the preserved text-layer identity conflict.

Phase 5 adds no temporal selection, selector attribution, runtime data, profile, manifest entry, package export, facade method, REST route, prediction, prognosis, diagnosis, advice, condition, score, or medical-truth claim. Phase 6 may now harden and expose inspectability over these private relation vessels, but inherits no resolved endpoint, direction, meaning, or scoring doctrine.

Constitutional Phase 4 Doctrine/Policy Closure — 2026-07-21

The Phase 4 decision has canonical SHA-256 4a444c91bab9a4949664e6bca4e64ad0ee341b439019db831429e4548bd2c4f9 and binds the exact Phase 3 decision at 2fd93585f8d2d439882ee77cdeb28e5509e916cd752357d60caaa003cc9fb2ca.

Phase 4 adds one private policy: moira_explicit_uromarisi_activity_ordinal_lookup_research_v1. Its vessel states that the caller supplies activity and ordinal, no temporal selector is source-attested or admitted, Stage 2K composition is unadmitted, interpretation is unperformed, medical use is forbidden, and admission remains research-only. The policy is a modern Moira control over the source-owned ordinal axis, not a computational rule attributed to Uromarisi.

The policy-governed operation requires all four inputs—corpus, policy, activity, and ordinal—and delegates exactly to Phase 3 lookup. It accepts no time or selector input and preserves absence without fallback. Machine cases bind EAT/1 to verse 230, RULE/4 to 244, DIE/5 to 255, and SLEEP/5 to no classification. Verse 250 remains independently visible only as a conflict.

The two admitted Stage 2K Bogamuni selector IDs are recorded as separately sourced, unbound candidates. No automatic composition or Uromarisi attribution is made. Phase 4 adds no source atom, relation doctrine, condition, score, prognosis, advice, runtime profile, manifest entry, export, facade method, or REST route. Phase 5 may now formalize relations, but inherits no resolved meaning for the activity-relation clauses.

Constitutional Phase 3 Inspectability Closure — 2026-07-21

The Phase 3 decision has canonical SHA-256 2fd93585f8d2d439882ee77cdeb28e5509e916cd752357d60caaa003cc9fb2ca and binds the exact Phase 2 closure at a5cd64696d4c040554f2c235056dfd28477fd0796fc82306f44ae43473d434e2.

Phase 3 adds derived-only views to the private cell, conflict, and complete corpus vessels. Cell views return stored identity, source binding, lexical marker names, mortality-language presence, and time-shape presence. Conflict views return the three stored text layers without precedence. Corpus views and typed lookups expose exact verses, counts, bindings, subsets, and identities without reconstructing source doctrine.

The absence policy is machine-checked: SLEEP ordinal 5 and verse 250 have no classification, while verse 250 has one separately retrievable conflict. There is no symmetry repair, adjacent lookup, DIE substitution, or default. Hardening rejects mutable containers, a wrong verse/ordinal matrix, activity-incompatible time classes, multiple source bindings within one activity, a conflict detached from Stage 2T, and malformed lookup inputs.

Phase 3 is complete at the private research boundary. It adds no semantic atom, classification, doctrine, condition, score, prognosis, advice, selector, profile, manifest entry, public export, facade method, or REST route. Phase 4 may now decide an explicit doctrine or policy, but it inherits no default, Uromarisi selector attribution, runtime outcome admission, or public API authorization.

Constitutional Phase 2 Classification Closure — 2026-07-21

The Phase 2 closure decision has canonical SHA-256 a5cd64696d4c040554f2c235056dfd28477fd0796fc82306f44ae43473d434e2. It binds the exact Stage 2Q–2U decisions and unchanged manifest rather than copying source expression into runtime data.

The private, data-free module moira._pancha_pakshi_classification supplies typed historical disposition, time-expression, and semantic-presence classes; immutable cell and text-layer-conflict vessels; and one complete-corpus vessel. The closure fixture records 24 classified cells with exact activity counts 5/5/5/4/5. Its tests locate the corresponding prior cell and independently derive the disposition, time class, semantic markers, source decision/hash, and uncertainty count before constructing the typed vessel.

Verse 250 remains outside classification. Its heading and verse say DIE ordinal 5 while its commentary says SLEEP ordinal 5, so it is represented only as a conflict with a null classification payload. The corpus rejects any attempt to classify that verse, lose a cell, duplicate an identity, change deterministic order, omit uncertainty, or mismatch mortality and disposition semantics.

Phase 2 is therefore complete at the private research boundary. Phase 3 may begin with derived inspectability and malformed-vessel hardening only. It may not create doctrine, interpretation, prognosis, advice, condition, score, selector binding, profile admission, package exports, facade methods, or REST routes. Temporal policy remains Phase 4. A competent Tamil reviewer is not a gate; Commissionerate A5 content is required only for a future independent corroboration claim; and the verse 256+ illness-duration section is a separate product.

Stage 2U Uromarisi DIE Semantics — 2026-07-21

Stage 2U reviews the five unconflicted DIE ordinals under the bounded semantic method. The Stage 2U research decision has canonical SHA-256 4954c13c33aa755bc0e8c6f47b7825d6ddb8346a2b6df39edf804872e81cbf70 and binds the unchanged Stage 2T, Stage 2P, and manifest hashes.

Rendered PDF pages 124–126, printed pages 122–124, control the reading; Archive.org OCR lines 5862–5964 provide navigation alignment only. Verses 251–255 produce five DIE semantic records. The first contains a rule-enmity mortality branch within two months and a distinct six-month tanmai branch that remains unnormalized. The second contains both an instantaneous marker and a one-year marker; no relation between them is invented. The remaining three cells state no numeric duration.

Verse 250 remains the blocked Stage 2T text-layer identity conflict. Stage 2U does not use it as a DIE cell, does not create a sixth ordinal, and does not resolve its SLEEP-versus-DIE disagreement. Verse 256 explicitly begins the separate illness-duration section and is excluded.

Mortality, life departure, body destruction, nonresolution, space/void, fate, deity, relation, and source-branch language remains historical and cell-local. It is not prediction, medical truth, diagnosis, prognosis, cause, symptom, score, advice, deterministic fate, calendar doctrine, or runtime relation. Tamil source expression, OCR, page images, and full translations are not distributed. No Stage 2K selector is attributed to Uromarisi, no Stage 2O result is routed to these records, and no profile, manifest, engine, facade, or REST surface changes. No human-language reviewer is required.

Stage 2T Uromarisi SLEEP Semantics and Identity Conflict — 2026-07-21

Stage 2T reviews the SLEEP locator sequence under the bounded semantic method. The Stage 2T research decision has canonical SHA-256 09f7651325cdac058d9816b85b031ef528f514ae91fb8cd9636452b8d7fb302a and binds the unchanged Stage 2S, Stage 2P, and manifest hashes.

Rendered PDF pages 122–124, printed pages 120–122, control the reading; Archive.org OCR lines 5743–5859 provide navigation alignment only. Verses 246–249 produce four unconflicted SLEEP records: within 8 days, within 15 days, exact 20 days, and a conditional mortality-or-resolution branch within three months.

Verse 250 is not an unconflicted fifth SLEEP cell. Its heading and verse assign DIE ordinal 5, while its commentary assigns SLEEP ordinal 5. Stage 2T preserves all three layers, performs no payload normalization, and leaves the prior hash-bound Stage 2P locator fixture unchanged. Current truth is four unconflicted SLEEP cells plus one blocked identity-conflict candidate.

Resolution difficulty, recurrence, mortality, distress, harm, wind-dosha, treatment, ritual, and relation language remains historical and cell-local. It is not prediction, medical truth, diagnosis, cause, symptom, score, advice, or runtime doctrine. Tamil source expression, OCR, page images, and full translations are not distributed. No Stage 2K selector is attributed to Uromarisi, no Stage 2O result is routed to these records, and no profile, manifest, engine, facade, or REST surface changes. No human-language reviewer is required.

Stage 2S Uromarisi RULE Semantic-Atom Extension — 2026-07-21

Stage 2S extends the bounded method to RULE ordinals one through five. The Stage 2S research decision has canonical SHA-256 85142480188a00ddec3de6f192a36025f282ca0eefa4643a6f1d74da4cec811d and binds the unchanged Stage 2R and manifest hashes.

Rendered PDF pages 120–122, printed pages 118–120, control the reading of verses 241–245; Archive.org OCR lines 5629–5740 provide navigation alignment only. The five source-stated time expressions are 3 days, 5 days, within 8 days, 10 days, and 12 days. All five cells state resolution, but the third remains an upper bound rather than an exact day.

Stage 2S adds only text-required atoms: source deity titles and their stated relations, explicit actions, fire-clause roles, Saturn-dosha references, bounded historical effect language, and a surface no-enmity statement in the fourth cell. Fire and dosha clauses are not diagnoses or medical causes, effect language is not converted to symptoms or scores, and the relation statement is not bound to runtime doctrine.

These are machine-assisted records of historical statements, not medical truth, diagnosis, prognosis, advice, treatment, condition scoring, electional judgment, or forecasting. Tamil source expression, OCR, page images, and full translations are not distributed. No Stage 2K selector is attributed to Uromarisi, no Stage 2O result is routed to these cells, and no profile, manifest, engine, facade, or REST surface changes. No human-language reviewer is required.

Stage 2R Uromarisi WALK Semantic-Atom Extension — 2026-07-21

Stage 2R extends the bounded Stage 2Q transcription method to WALK ordinals one through five. The Stage 2R research decision has canonical SHA-256 361a0a334a73623cb0b2c1b0e73489db2c20d3c259e04540a303510113f0e0d6 and binds the unchanged Stage 2Q and manifest hashes.

Rendered PDF pages 118–120, printed pages 116–118, control the reading of verses 235–239; Archive.org OCR lines 5488–5603 provide navigation alignment only. The five time expressions are 10 days, 15 days, within 20 days, 25 days, and within one month. Exact and upper-bound forms remain distinct, and the month is not converted into days.

The source does not use one uniform disposition. Ordinals one and four state resolution, ordinals two and three state abatement, and ordinal five describes a timed progression without explicitly stating resolution. Stage 2R therefore extends rather than flattens the Stage 2Q ontology. Source deity titles and their stated relations, explicit responses, medicine or physician references, water-clause roles, the Navagraha-dosha reference, unresolved activity relations, and uncertainty remain cell-local.

These are machine-assisted records of historical statements, not medical truth, diagnosis, prognosis, advice, treatment, condition scoring, electional judgment, or forecasting. Tamil source expression, OCR, page images, and full translations are not distributed. No Stage 2K selector is attributed to Uromarisi, no Stage 2O result is routed to these cells, and no profile, manifest, engine, facade, or REST surface changes. No human-language reviewer is required.

Stage 2Q Uromarisi EAT Semantic-Atom Pilot — 2026-07-21

Stage 2Q continues from the complete Stage 2P locator grid without admitting an outcome product. The Stage 2Q research decision has canonical SHA-256 7b4311912ece7f49b30773604c91537ca5fa2a9e02b75baeebfb5bdc2575bcd9 and binds the unchanged Stage 2P and manifest hashes.

The pilot covers only EAT ordinals one through five, verses 230–234 on rendered PDF pages 116–118. Rendered-page review controls the semantic reading; Archive.org OCR lines 5380–5486 supply navigation alignment only. The second and third durations are seven and nine days respectively, correcting an earlier visual-only reading before any fixture was admitted. The resulting source-stated duration sequence is 4 or 5, 7, 9, 13, and 15 days.

The semantic vocabulary remains cell-local and source-owned: stated resolution and duration, distinct devotional-response categories, medicine and prithivi term references, unresolved relation clauses, and explicit uncertainty. The repeated relation clause is recorded as present but retains null semantics. Named offering details, Tamil source expression, OCR, page images, and full translations are not distributed.

These records summarize what one historical witness states. They do not establish medical truth and are not diagnosis, prognosis, advice, treatment, condition scoring, electional judgment, or forecasting. No Stage 2K selector is attributed to Uromarisi, no Stage 2O result is routed to these cells, and no profile, manifest, engine, facade, or REST surface changes. The pilot is machine-assisted and carries no human-language reviewer dependency.

Stage 2P Uromarisi Illness-Grid Research Gate — 2026-07-21

Stage 2P advances the remaining Uromarisi branch by recovering one provisional locator structure without translating or admitting its prognostic payload. The Stage 2P research decision has canonical SHA-256 449efb11b81741e1ac591d6a93033023f67892ac835cbcb178103606eb729dd2. It binds the unchanged Stage 2J research decision, Stage 2O decision, and manifest SHA-256 values.

The exact Archive.org 1934 Uromarisi-attributed PDF was downloaded and locally verified at SHA-256 e2ab7a64d4d4e540c30bc464c12923e6f14e93fbbe15d73e459e9c62a5815da0. Rendered original-page review covered PDF pages 115–126, corresponding to printed pages 113–124. The provisional 25-locator index binds verses 230–239 and 241–255 to their PDF and printed-page spans. Stage 2T later establishes that verse 250 has a heading/verse versus commentary parent- activity conflict and is not an unconflicted fifth SLEEP cell. Verse 240 is an intervening transition and is not an ordinal cell. Verse 256 begins a separate illness-duration section after the grid.

This is structural source evidence only. The decision preserves activity, ordinal, verse, and page locators but copies no Tamil source expression and performs no translation or outcome normalization. The grid states no selector arithmetic, so neither Bogamuni Stage 2K policy is attributed to Uromarisi and no Stage 2O civil-time result is routed into it. Medical advice, diagnosis, prognosis, condition mapping, numeric scoring, electional judgment, and forecasting remain outside the admitted surface. The four profiles, manifest, engine, facade, and REST routes are unchanged. No human-language reviewer is required.

Stage 2O Explicit Civil-Time-To-Sookshma Routing — 2026-07-21

Stage 2O binds the existing timing materializations to the Stage 2N composition under a new, explicitly modern routing policy. The Stage 2O decision has canonical SHA-256 2ea686e774ba4468c0515f621771b8a142c79f04d89b69839f482e05c37b40df. It binds the unchanged Stage 2N decision SHA-256 084190606dc358abce7cc1879aa898a0071bce421b1eda8845b113520a7c36a9 and unchanged manifest SHA-256 584d2b28bd2c7537f8ebb029633ed7bce682ed02ee38bf32402901940887c955.

Every call names both source profiles, an aware civil instant and location, source Paksha, subject bird, one of the existing fixed-clock or solar-proportional materialization policy IDs, and one Stage 2K selector policy. The current materialized cell supplies the samam. The stored reader-bound TT requested instant and samam endpoints are lifted from binary64 exactly into rational arithmetic and normalized to the samam’s six nazhigai span before the unchanged Stage 2N join is evaluated. Shared boundaries belong to the following samam.

The policy is civil_time_materialized_samam_to_stage2n_v1 and modern_moira_policy_not_source_claim. It neither attributes this routing to a historical witness nor chooses a timing or selector default. A fixed-clock long-half tail remains unmaterialized_solar_half_tail with no samam, elapsed offset, or composition; no proportional fallback occurs. Paksha remains caller supplied. No astronomical paksha inference, Uromarisi outcome, condition, score, electional judgment, or forecast is produced. All four profiles, their capabilities, and the manifest remain unchanged, and no human-language reviewer is required.

Stage 2N Explicit Schedule-To-Sookshma Composition — 2026-07-21

Stage 2N resumes the implementation sequence while leaving Commissionerate A5 as optional future corroboration. The Stage 2N composition decision has canonical SHA-256 084190606dc358abce7cc1879aa898a0071bce421b1eda8845b113520a7c36a9. It binds the unchanged 1879 schedule profile SHA-256 d80d205716eb9f24a2a23949c6df241a1aba251749efa94d3b20fa36be0258f4, unchanged 2024 selector profile SHA-256 596c003c62ebbda913ca28aef318d77cb7b1cf42d92d3b1b7a20a44a01dd6526, and unchanged current manifest SHA-256 584d2b28bd2c7537f8ebb029633ed7bce682ed02ee38bf32402901940887c955.

The governing object is a caller-named schedule samam and subject bird joined to one explicitly selected Stage 2K policy at an exact elapsed samam offset. The operation generates the named schedule, requires the subject bird’s unique cell in samam 1..5, and passes that cell’s activity to the unchanged Sookshma selector. Every input remains mandatory: both profile IDs, source Paksha, half, weekday, samam, subject bird, selector policy, and exact reduced Fraction. The returned vessel preserves the complete schedule and its 1879 provenance alongside the complete selector result and its 2024 provenance.

The join is explicitly explicit_schedule_samam_subject_bird_sookshma_v1, a modern_moira_policy_not_source_claim; it is not attributed to either source. It performs no datetime, civil-clock, astronomical, Uromarisi outcome, translation, condition, scoring, electional, or forecasting operation. The equal-fifths policy still supplies ordinals only and invents no subactivity. All four profile entries and their capabilities remain unchanged. No human Tamil reviewer or A5 manuscript access is required for this bounded modern composition.

Stage 2M Ramadevar Candidate Identity And Access Gate — 2026-07-21

Stage 2M follows the Ramadevar lead to the exact institutional record without turning a catalog title into computational evidence. The Stage 2M research decision has canonical SHA-256 921e604bcd81298aa6eb903acc967e68cfcf6e743c7d1379788ff9996212c6db. It binds the unchanged Stage 2L decision SHA-256 5534ddde1c0b87fa5fc3332112d02fd1c48c38e0a79f45f4a75a3e3c728a4c34 and manifest SHA-256 584d2b28bd2c7537f8ebb029633ed7bce682ed02ee38bf32402901940887c955.

The 63-page Catalogue of Palm Leaves Manuscripts at Commissionerate of Indian Medicine and Homoeopathy Department, Arumbakkam has SHA-256 8bf4541aa46e3526d3218b1c35ae7bf174298ff6e29e86d9e10e7389dd5b5e4b. Rendered pages 1 and 2 establish the holding collection and its broad medical, astrological, magical, yoga, and gnana scope. Rendered page 52 identifies serial 859, manuscript A5, as Ramadevar Panchapakshi. That row supplies no leaves, transcription, dimensions, condition, completeness, script, date, incipit, explicit, colophon, prior ownership, acquisition path, copying history, or computational rules. The candidate is real and exactly located, but remains title-only; neither product comparability nor textual independence can be assessed.

The search also resolves misleadingly similar Ramadevar works. The 1999 G.O.M.L. descriptive catalog, SHA-256 4ff7f72891c6d53c3eaac502f1f1217a0cb950b60611524bfaff854d38b03ec4, classifies complete manuscript R.8978 Ramadevar Patchini on rendered PDF pages 93-94 as a 21,648-letter work on gnana, breath, and yoga. British Library EAP EAP1217/1/2851, original Tamil University reference TU_TAMIL_2058-01_2661, is independently catalogued as an eighteenth-century 108-poem work on philosophy and ashtanga yoga. The Commissionerate catalog itself lists a different Ramadevar Patchani - 108 inside manuscript 27 on page 4. Finally, the 1991 Sarasvati Mahal Ramadevar Sutiram: Ashtanga Patchini, Valai Pujai, SHA-256 238987ee86aba6b17c963031616fc58598fa308df0c8ffec2a6314ed89156021, is framed on rendered pages 4, 7, and 10 as medicine, alchemy, yoga, ritual, and related Siddha practice. None supplies the five-bird temporal product.

Patchani is therefore not an automatic spelling alias for Panchapakshi, and shared Ramadevar attribution or a 108-poem count does not establish work identity. Clearing the gate now requires the actual Commissionerate A5 leaves or a critical transcription, item-level physical and copying-history metadata, and a leaf-located collation of one named Pancha Pakshi product. Stage 2M changes no runtime data, manifest, admission status, engine export, facade method, or REST route. No human Tamil reviewer is required.

Stage 2L Independent-Witness Collation Gate — 2026-07-21

Stage 2L completes the next research gate without admitting a new profile or raising an existing profile to corroborated_public. The Stage 2L research decision has canonical SHA-256 5534ddde1c0b87fa5fc3332112d02fd1c48c38e0a79f45f4a75a3e3c728a4c34. It binds the unchanged Stage 2K manifest SHA-256 584d2b28bd2c7537f8ebb029633ed7bce682ed02ee38bf32402901940887c955 and Stage 2K decision SHA-256 10bcfbd70dda28fd399e5c95b8bfa237b8e48f3b2cb20901fc21e0261a73cf70.

The strongest comparator is the Tamil Digital Library artifact for the Sarasvati Mahal Library 2014 sixth edition of Valaiyarul Patinen Siddhargal Panchapakshi Sastram. The inspected 574-page PDF has SHA-256 894f88c3381f026aa1963861dd30e1f74039aa32a89fd84015efb3a098dc5366. Rendered pages 2 and 4 establish the edition and institutional source method; pages 30, 31, 33, and 37 preserve the editor’s source-layer statements, including Agastya manuscript 777, Kumbamuni manuscript 7015, later Bogar and Uromarisi additions, and explicit disagreement between Agastya/Bogar and Uromarisi activity orders. Pages 88 and 89 preserve the Purva-day first-EAT rule and the four day/night ordering doctrines. Pages 136 and 209 preserve all four regime-specific duration vectors, and page 302 shows the editor’s manuscript-based Purva-day first-samam expansion.

The exact seven-cell Purva-day first-EAT mapping and waxing-day duration vector agree with the admitted 1879 profile. Three duration vectors do not: the institutional compilation owns separate waxing-night, waning-day, and waning-night values, while the 1879 profile owns one uniform vector. The supplied 89-page G. R. Narasimhan guide, SHA-256 1694ca303d5f49a13b9269fe9ee1e39607e1709df756ff6bbdf137bc1c589243, reproduces all four institutional vectors but contains no bibliography, citation, or primary-text lineage. It is therefore useful as a modern secondary comparator only. The supplied 21-page Canva artifact, SHA-256 3498501cf3fd43ec2b0f2b8dab497b6c431713af1373f229a496bb8c7f6b7b9d, is rejected: metadata marks AI-generated content and rendered pages 6 and 10 contain unrelated tarot-deck material.

The Sarasvati volume is a separately published institutional compilation, but its declared source overlap and lack of a manuscript stemma or copying history leave textual-lineage independence unresolved. Catalog records for a Ramadevar Panchapakshi palm-leaf manuscript and Raghavanandana’s Sanskrit Panchapakshi Tika identify future candidates, but their computational text was not available for collation. Publication separation, catalog existence, or exact agreement alone cannot clear the gate. Stage 2L therefore changes no runtime data, manifest, admission status, engine export, facade method, or REST route. No human Tamil reviewer is required.

Stage 2K Sookshma Temporal-Selector Admission — 2026-07-21

Stage 2K admits the two source-attested Bogamuni 2024 editorial selectors as explicit, separate policies on a new source-scoped profile. The Stage 2K admission decision has SHA-256 10bcfbd70dda28fd399e5c95b8bfa237b8e48f3b2cb20901fc21e0261a73cf70. It binds profile SHA-256 596c003c62ebbda913ca28aef318d77cb7b1cf42d92d3b1b7a20a44a01dd6526 and current manifest SHA-256 584d2b28bd2c7537f8ebb029633ed7bce682ed02ee38bf32402901940887c955.

The governing object is an exact elapsed rational offset within one six-nazhigai samam. Every call must name one of these policies:

Both policies use exact [start, end) ownership over [0, 6). Neither is a default, they are not interchangeable, and no automatic selection is allowed. The selected policy, all five intervals, unique ordinal and interval, source locators, and provenance remain visible. The admission performs no datetime, civil-clock, astronomical, schedule, Uromarisi-outcome, condition, score, electional, or forecasting composition. It creates no human-language-review dependency. The live 1879 profile also replaces its obsolete pending-review token with machine_reconciled_source_assignment_with_declared_uncertainty; frozen historical receipts retain their original evidence labels.

Stage 2J Vinadi Research Gate — 2026-07-21

Stage 2J is a research recovery, not a public admission. The Stage 2J research decision has canonical SHA-256 d04ed0f3716fe605dc5d8172114dc759b30c4e87be968eebc36e35a23d789243. It binds the unchanged Stage 2I manifest SHA-256 d1aba3757910ded019cb6a2a5d6fb92c2e1ebbea755c26953dff1347834bf0e8 and Stage 2I decision SHA-256 83c9bc0a423c09ccc113007625fee4a7d6b9ee1e890827f71595c96c3f826807. No runtime profile, data, manifest capability, engine vessel, facade method, or REST route changes.

Rendered inspection of the Uromarisi-attributed 1932 publication recovered a source-owned five-position vinadi ordinal axis beneath each of the five named activities. PDF page 5 enumerates first through fifth vinadi result headings under multiple contexts; PDF page 88 contains governing verses 228 and 229; PDF page 89 begins the ordinal result sequence. Independent digest re-verification corrected the previously recorded value: the inspected 197-page PDF has SHA-256 dbd12d7e26f39ca7f9650a17311b5483eb478844144544a2cbb11aac7c3d6243. The official Tamil Digital Library catalog identifies the 1932 Chennai publication and Sarasvati Mahal Library holding.

The official 1922 Tamil Digital Library publication includes Karunanidhi Pillai commentary. Rendered PDF page 115 (printed page 113) preserves verses 227 through 229; page 116 (printed page 114) restates the demand for accurate nazhigai division but supplies no selector arithmetic. The inspected 298-page PDF has SHA-256 51b4b34890412fd57011aebe0c1ab22ab1800e5035a84bbbb9330ea0f6597741.

The separately published Uromarisi-attributed 1934 volume preserves the same verse numbers and ordinal organization. Rendered inspection located verses 227 through 229 at PDF page 115 (printed page 113) and the first through fourth EAT-vinadi outcomes beginning at PDF pages 116 through 117 (printed pages 114 through 115). Its inspected 166-page PDF has SHA-256 e2ab7a64d4d4e540c30bc464c12923e6f14e93fbbe15d73e459e9c62a5815da0. This is separate-publication corroboration; textual-lineage independence is not established.

The Uromarisi witnesses require accurate vinadi placement but do not supply selector arithmetic on the inspected governing pages. The separately named editorial layers in the Bogamuni-attributed 2024 publication do supply two incompatible selector candidates. Rendered PDF pages 157 and 158 (printed pages 147 and 148) give weighted Sūkṣma durations EAT 3/2, WALK 5/4, RULE 2, SLEEP 3/4, and DIE 1/2 nazhigai, closing exactly to one six-nazhigai samam, with cyclic rows beginning at the named activity. PDF page 169 (printed page 159) separately names Eka Sūkṣma and directs that each samam’s nazhigai be divided into five equal parts. The inspected PDF has SHA-256 035eab41f62cf078180c03e99ec9eacf8edf2d2dc6d3dc31b37e6a6dfdb09990.

The two doctrines are preserved as provisional policies bogamuni_2024_weighted_sookshma_samam_v1 and bogamuni_2024_eka_sookshma_equal_fifths_v1. They are not interchangeable, neither is a default, and neither is automatically bound to Uromarisi ordinal outcomes. Machine-assisted reading is insufficient to normalize prognostic prose, but there is no human-language-review dependency. Stage 2J therefore remains research-only because selector choice and cross-witness composition are unadmitted. Automatic temporal routing, translation-backed outcomes, composition with admitted schedules or identities, condition, scoring, electional judgment, and forecasting remain forbidden pending their own explicit normalization and admission.

Stage 2I First-Samam EAT-Seed Addendum

Stage 2I admits first_eat_bird_mapping as one granular capability of the unchanged agastya_madras_1879_akshara_fixed_clock profile. It does not add a fourth profile or alter the normalized 1879 source data. The Stage 2I admission decision has canonical SHA-256 83c9bc0a423c09ccc113007625fee4a7d6b9ee1e890827f71595c96c3f826807. It binds the unchanged profile SHA-256 4fe769b6f13c4a719c9d31446dd3fef413eca5d3ce1f56340aada9f99b0dce64, the frozen Stage 2H decision SHA-256 9ea7c871643bb8fc68d420223d0090ca91699154c761c67ccaf9201f401906cd, and the manifest transition from eae9fc471da08eccf24515ef12cdaf59330aa1b7ad7f9d43432c7a1482704a03 to d1aba3757910ded019cb6a2a5d6fb92c2e1ebbea755c26953dff1347834bf0e8.

The governing witness remains the Agastya-attributed Madras 1879 print. Its Internet Archive derivative PDF is dli.rmrl.000451.pdf, archive MD5 0736b952fb587132c2181a383ff10cfb, archive SHA-1 d41ff5c2d569de6422435b20135b58be82a68560, and locally verified SHA-256 ed52945ee141faa3f6967b8f043077b95abef9ff674ffb83eaba633417c669c9. Rendered-page inspection, with OCR used only for navigation, binds the first samam’s weekday EAT seed at Purva-day leaf n16, Purva-night leaf n21, Amara-day leaf n26, and Amara-night leaf n31.

Context Sunday Monday Tuesday Wednesday Thursday Friday Saturday
Purva day Vulture Owl Vulture Owl Crow Cock Peacock
Purva night Crow Cock Crow Cock Peacock Vulture Owl
Amara day Cock Peacock Cock Crow Owl Vulture Peacock
Amara night Vulture Cock Vulture Owl Crow Peacock Cock

These 28 values are schedule-generator seeds, not ambient or whole-day eating birds. Each seed names the bird that begins the selected generator’s first samam in EAT; later samams advance under the separately normalized progression rule. The public result therefore carries generator_id, exact source axes, first_eat_bird, the complete canonical generator locator tuple, and profile provenance without materializing a schedule.

The 1934 Uromarisi publication provides separate-publication corroboration of all 28 cells at leaf n6 and leaves n36-n37. The inspected PDF’s SHA-256 is e2ab7a64d4d4e540c30bc464c12923e6f14e93fbbe15d73e459e9c62a5815da0. This is a physically separate publication, but independent textual lineage is not established; it does not contribute runtime cells or raise the profile to corroborated_public. The frozen blind reading is also not an independent witness and is superseded for Amara-night cells by the later n31-n35 adjudication reflected in the canonical profile.

The public surfaces are the kernel-free pancha_pakshi_first_eat_bird_mapping(...), Moira.pancha_pakshi_first_eat_bird_mapping(...), and strict POST /v1/pancha-pakshi/schedule/first-eat-bird. They accept only explicit profile_id, profile_paksha, day/night half, and weekday. They perform no datetime or location routing, astronomical Paksha inference, schedule materialization, natal identity, or cross-profile composition. The result is not Padu, an authority day or bird, Adhikara/Bharana, a current activity, condition, strength, score, electional judgment, or forecast.

The current public family has four profiles, nineteen Moira methods (eleven kernel-free and eight kernel-backed), and nineteen REST routes.

Validation checks all 28 cells against the canonical generator table and the existing schedule’s first_eat_bird and initial EAT cell. It also checks exact locator/provenance binding, capability isolation, vessel immutability, strict facade/REST/OpenAPI shapes, and reconstruction of the frozen Stage 2F and 2H manifests. No source PDF, OCR, rendered page, copied layout, source prose, or translation is bundled.

Stage 2H Padu-Bird Mapping Addendum

Stage 2H admits the separate bogamuni_chennai_2024_padu_bird_mapping profile with product kind and sole capability padu_bird_mapping, source_scoped_public status, and no default. The Stage 2H admission decision has canonical SHA-256 9ea7c871643bb8fc68d420223d0090ca91699154c761c67ccaf9201f401906cd. It binds profile SHA-256 5de0d1e28d47fad8be6a2a1ab648f2ed71eaf742be2775d166ea44981e96ff10 and manifest SHA-256 eae9fc471da08eccf24515ef12cdaf59330aa1b7ad7f9d43432c7a1482704a03, while freezing the prior Stage 2G decision at 3da998bc78c6c1fc4ec1c71629dc6b3872725ef25f73965474d5e894deec1575.

The primary witness is the same Bogamuni-attributed 2024 sixth edition used as a separately scoped source in Stage 2G. The Internet Archive original PDF is Acc.No.44757-PanjapatchiSashthiram-2024.pdf, archive MD5 abe489a832ac38a0270335b7429776f3, archive SHA-1 6ddad8f2577883f6859829f534e8ee7b8330ade8, and locally verified SHA-256 035eab41f62cf078180c03e99ec9eacf8edf2d2dc6d3dc31b37e6a6dfdb09990. Rendered original-page inspection, with OCR used only for navigation, found the governing Purva weekday Padu stanza and commentary at IA leaf n52, the Amara material at n60, the repeated combined table at n157, and its restating commentary at n158.

Stage 2H declares paksha_stanzas_govern_repeated_combined_table_confirms. The Paksha stanzas govern; the combined table and commentary repeat and confirm the same fourteen cells. The normalized table is exactly two explicit source Paksha labels by seven weekdays:

Paksha Sunday Monday Tuesday Wednesday Thursday Friday Saturday
Purva Owl Crow Cock Peacock Vulture Owl Vulture
Amara Crow Owl Vulture Peacock Cock Peacock Cock

There is no day/night axis. PanchaPakshiPaduBirdMapping preserves Padu as the weekday’s death-or-inoperative bird and explicitly states that it is not the schedule’s RULE activity. The source layers separately label an eating-bird table and authority days; they do not present an Adhikara Pakshi table. Bharana is secondary-only terminology. Moira therefore admits no Adhikara/Bharana alias or product and does not relabel the existing first_eat_bird schedule field. Uromarisi 1934 and the separately inspected Bogar material are not bound by the Stage 2H Padu profile or decision; they contribute no Padu runtime cell. Stage 2I later records Uromarisi only as corroboration of the separate 1879 first-EAT object.

The public surfaces are the kernel-free pancha_pakshi_padu_bird_mapping(...), Moira.pancha_pakshi_padu_bird_mapping(...), and strict POST /v1/pancha-pakshi/roles/padu. They accept only explicit profile_id, profile_paksha, and weekday. The result exposes the selected bird, direct attestation status, source semantics, assembly policy, three canonical locators, provenance, and omissions. It performs no astronomical or civil-day routing, natal identity, schedule/materialization/current-cell operation, condition evaluation, score, or forecast. At the Stage 2H boundary, the public family had three profiles, fourteen Moira methods (seven kernel-free and seven kernel-backed), and thirteen REST routes.

Validation keeps evidence classes separate. The four rendered pages and exact fourteen-cell projection are source-table evidence. Completeness, Paksha/weekday ownership, no-day/night shape, canonical locator binding, immutable provenance, capability separation, and strict facade/REST/OpenAPI contracts are structural evidence. Separately observed Uromarisi/Bogar material is not fixture-bound and does not make the runtime table cross-witness or universal, and there is no external condition, scoring, electional, or forecasting oracle. As in every prior phase, no source PDF, OCR, rendered page, copied layout, source prose, or translation is bundled.

Stage 2G Natal-Moon Identity Addendum

Stage 2G admits the separate bogamuni_chennai_2024_nakshatra_natal_identity profile with product kind natal_moon_bird_identity, capabilities nakshatra_bird_mapping and natal_identity, source_scoped_public status, and no default. The Stage 2G admission decision has canonical SHA-256 3da998bc78c6c1fc4ec1c71629dc6b3872725ef25f73965474d5e894deec1575. It binds profile SHA-256 e3642f61756ed7b8c413ddfbde2844769aea1994d4e69ae27594b2059b549b6a and manifest SHA-256 979bb6df8a31d0ff9603ef396b0f569f17ecca6f6dc21def220ad682a425eb61.

The primary witness is the Bogamuni-attributed 2024 sixth edition catalogued by Internet Archive as acc.-no.-44757-panjapatchi-sashthiram-2024. The inspected original PDF is Acc.No.44757-PanjapatchiSashthiram-2024.pdf, archive MD5 abe489a832ac38a0270335b7429776f3, archive SHA-1 6ddad8f2577883f6859829f534e8ee7b8330ade8, and locally verified SHA-256 035eab41f62cf078180c03e99ec9eacf8edf2d2dc6d3dc31b37e6a6dfdb09990. Rendered original-page inspection—not OCR alone—established the complete Purva nakshatra-bird table at IA leaf n52, the complete Amara verse at n64, and the source phase-to-Purva/Amara binding at n167.

The commentary adjacent to the Amara verse is internally malformed: it duplicates Shravana and omits Revati. Stage 2G declares verse_precedence_for_nakshatra_partition, uses the complete verse for the 27-cell Amara partition, and preserves the commentary at n64 as rejected conflict evidence. It does not infer the omitted cell, repair the commentary by symmetry, or blend text layers. The Uromarisi-attributed 1934 witness at kvc-0354-vinaadi-pajasapatchi-mulamum-1934 corroborates the Purva grouping at leaf n18 and independently exhibits malformed Amara commentary at n61. That bounded corroboration does not govern the runtime table or raise the entire profile to corroborated_public.

PanchaPakshiNakshatraBirdMapping preserves the pure 2-by-27 source table and states nakshatra_bird_table_not_explicitly_natal_moon. The source witnesses attest phase labels and nakshatra birds; they do not explicitly prescribe a birth-Moon calculation, Lahiri ayanamsa, or an equal-27-sector numerical taxonomy. PanchaPakshiNatalMoonIdentityPolicy therefore exposes composition_status="modern_moira_policy_not_source_claim" and the sole policy ID bogamuni_2024_apparent_lahiri_natal_moon_identity_v1.

The modern composition converts one explicit UT1 instant once to reader-bound TT. Apparent geocentric Sun and Moon longitudes in the true ecliptic of date determine the half-open Shukla/Krishna phase and source Paksha. The same TT epoch governs Lahiri true ayanamsa and the sidereal Moon. Twenty-seven equal half-open 40/3-degree sectors assign exact internal boundaries to the following nakshatra. A maximum-one-ULP-below recovery exists only to recover the binary representation of an exact mathematical boundary and is not a tolerance band.

The public surfaces are the kernel-free pure mapping function and facade method, the kernel-backed pancha_pakshi_natal_moon_identity_at(...) and Moira.pancha_pakshi_natal_moon_identity(...), and strict POST /v1/pancha-pakshi/identity/natal-moon. REST accepts only profile_id, aware dt, and the exact policy ID. It rejects location, supplied paksha/nakshatra/bird/ayanamsa, schedule/current-cell controls, scoring, and forecast inputs. The result exposes every astronomical and sidereal intermediate, phase and bird mappings, locators, modern composition status, provenance, and omissions; no schedule routing occurs.

Validation keeps its evidence classes separate. The named pages and exact 54-cell projection are source-table evidence. Exact-boundary and adjacent-representable tests are mathematical invariants. DE441 executes the apparent geocentric reader-bound TT path, but is not a Pancha Pakshi or natal-identity oracle. Facade, serializer, REST, OpenAPI, export, profile/hash, capability, and immutable-vessel tests protect the public contract. Exact source locators and reading uncertainty remain visible; there is no human-language-review dependency for this narrow source-scoped admission.

The source-artifact boundary remains unchanged. The PDFs and rendered research pages are not bundled; Moira distributes only independently normalized symbolic rules, source identities and locators, Moira-authored policy/schema/prose, and validation metadata. The 1879 aksara profile remains distinct and unchanged.

Stage 2F Astronomical Paksha Inference Addendum

Stage 2F admits astronomical_paksha_inference for the same explicitly named source_scoped_public profile. Unlike Stages 2A-E, it adds a normalized source fact to the profile: machine-assisted visual reading of the exact 1879 witness directly maps waxing to Purva at IA leaf n16 and waning to Amara at n26. The frozen source-reading record has canonical SHA-256 9ce3686a90a41af916a370b8d4ec04637f22a1d32f872180c6d8a1b790e25a0e. Its status remains machine_assisted_visual_reading_pending_competent_tamil_review; this legacy provenance label records how the source reading was originally classified and does not create a human-review dependency. It does not claim independent- witness corroboration or a universal Pancha Pakshi vocabulary.

The chained Stage 2F admission decision has canonical SHA-256 1020b28d5da8d0e823cadd352ea2236c69cbb636660a573eb5d74b8c131bc5d8. It freezes the Stage 2E decision at 4ddf0a5fa5b680fa83a7bb3052ecbc5d1a9c2f685c466290f22121dd02724d18, the prior manifest at d2b5f8f1ae7e067d257eeb24b533be1d33349446d56d361ea59f4a71472eca70, and the prior profile at 876e4cc7cc5d894f5e558ac733913e84a8b779f72c77661e89d448fd1e05ced4. The schema-3 profile digest becomes 4fe769b6f13c4a719c9d31446dd3fef413eca5d3ce1f56340aada9f99b0dce64 and the schema-2 manifest digest becomes a4fdceee4089c2812d9d77be763c1738152a63231b3f06847ea93383e4a3b327. Admission status, product kind, and default_selection_allowed=false remain unchanged.

The governing objects are PanchaPakshiAstronomicalPakshaInference, the PanchaPakshiAstronomicalPaksha enum, and immutable PanchaPakshiAstronomicalPakshaInferencePolicy. The only admitted policy is apparent_geocentric_moon_sun_longitude_paksha_half_open_v1. It converts the explicit UT1 instant once to reader-bound TT and evaluates apparent geocentric Sun and Moon longitudes in the true ecliptic of date on that same coordinate. Normalized Moon - Sun longitude assigns [0, 180) to Shukla/waxing/Purva and [180, 360) to Krishna/waning/Amara. Exact 0 belongs to Shukla/Purva and exact 180 to Krishna/Amara; tolerance, snapping, civil-day override, and topocentric override are absent. Ayanamsa has no role because a common longitude offset cancels from the phase difference.

The low-level engine, facade, and REST surfaces are pancha_pakshi_astronomical_paksha_at(...), Moira.pancha_pakshi_astronomical_paksha(...), and POST /v1/pancha-pakshi/context/astronomical-paksha. The transport request accepts only explicit profile_id, aware dt, and the exact policy ID. There is no location or caller-supplied paksha. The result publishes UT1/TT, both longitudes, normalized elongation, astronomical half, source-mapped profile label, exactly one direct mapping locator, policy, and provenance. It does not select or materialize a schedule, choose a current cell, route its result into another operation, or infer natal identity.

Stage 2F validation is intentionally partitioned. The leaf readings establish only the source-label mapping. Synthetic 0/180 and adjacent-representable cases establish the modern half-open classifier; DE441 exercises the actual reader-bound astronomical path; and Panchanga coherence checks confirm that tithi and karana use the direct tropical Moon-Sun difference, where common ayanamsa cancels, rather than separately rounded sidereal operands. Facade, transport, OpenAPI, capability, immutable-vessel, schema/hash, and locator checks protect the public contract. None of this is an external Pancha Pakshi oracle, a new phase-event timing accuracy claim, linguistic-authority claim, or independent-witness collation.

The standing source-artifact boundary is unchanged. Stage 2F distributes only normalized facts, locators, Moira-authored policy, and validation metadata. It does not package an archival scan, PDF, OCR, page image, copied source expression, or third-party translation.

Stage 2E Solar-Proportional Current-Cell Addendum

Stage 2E admits one further capability, solar_proportional_current_cell_selection, for the same source_scoped_public profile. The profile document remains unchanged at canonical SHA-256 876e4cc7cc5d894f5e558ac733913e84a8b779f72c77661e89d448fd1e05ced4, and default selection remains forbidden. The chained decision is recorded in pancha_pakshi_1879_solar_proportional_current_cell_2026_07_20.json, whose canonical SHA-256 is 4ddf0a5fa5b680fa83a7bb3052ecbc5d1a9c2f685c466290f22121dd02724d18. It freezes the Stage 2D decision at SHA-256 e31e0664090b9a38bdcd52b660c04998a0412eab223f4a84fe745b9e54d25383, the prior manifest at SHA-256 6dbbf05383c7a4eb3eadebf70fdb1130ab5081ef83175bc387755ffda4db9121, and binds the new manifest at SHA-256 d2b5f8f1ae7e067d257eeb24b533be1d33349446d56d361ea59f4a71472eca70.

The governing object is PanchaPakshiSolarProportionalCurrentCellSelection, with immutable PanchaPakshiSolarProportionalCurrentCellSelectionPolicy. Its only admitted policy is solar_proportional_current_cell_half_open_solar_precedence_v1, explicitly a modern deterministic Moira selection policy rather than an 1879 source claim.

Stage 2A first resolves the governing half-open solar half. Stage 2D then materializes that half’s complete 25-cell proportional partition with the same configured reader. The requested UT1 instant is converted once to reader-bound TT and tested by exact start_jd_tt <= requested_jd_tt < end_jd_tt membership with 0.0 s tolerance. The anchor belongs to cell zero, every shared endpoint belongs to the following cell, and exact sunrise or sunset is routed into the new governing half before selection.

Complete Stage 2D coverage makes selected the only lawful status and requires one non-null tuple member. Zero or multiple matches fail closed. Stage 2E has no fixed-clock tail state and performs no clipping, wrapping, repetition, borrowing, fixed-clock mixing, fallback, paksha inference, or natal identity. Paksha remains an explicit caller-supplied source label.

The low-level engine, facade, and REST surfaces are respectively pancha_pakshi_solar_proportional_current_cell_at(...), Moira.pancha_pakshi_solar_proportional_current_cell(...), and POST /v1/pancha-pakshi/schedule/solar-proportional/current-cell. The compact REST response publishes one non-null cell and governing bounds without duplicating the complete Stage 2D materialization. Result provenance equals the Stage 2D provenance with only its routing status changed to solar_proportional_current_cell_selection_performed_paksha_caller_supplied_no_fixed_clock_mixing_or_inference; the omission source_attested_solar_proportional_materialization remains unchanged.

Validation is bounded to solar-half-first routing, exact TT interval ownership, complete-coverage and exactly-one-match invariants, policy/result immutability, provenance identity, admission/hash integrity, facade/transport strictness, and failure of gap, overlap, foreign-cell, tail-state, or mixed-policy vessels. The configured content-identified DE-0441LE-0441 reader exercises the published UT1-to-TT boundary path. This is not an external Pancha Pakshi current-cell oracle, independent-witness corroboration, or a new historical-accuracy claim.

Stage 2E Implementation Verification

The admitted implementation was verified on 2026-07-20 with the project Python 3.14.3 .venv, MOIRA_TEST_MODE=1, MOIRA_STRICT_KNOWN_ISSUES=1, and downloads disabled:

This receipt validates the deterministic policy, public contract, and reader-bound numerical behavior. It does not turn DE441, structural interval invariants, or the inherited Horizons solar-boundary fixture into an external Pancha Pakshi current-cell oracle.

Stage 2D Solar-Proportional Materialization Addendum

Stage 2D admits one further capability, solar_proportional_materialization, for the same source_scoped_public profile. The profile document remains unchanged at canonical SHA-256 876e4cc7cc5d894f5e558ac733913e84a8b779f72c77661e89d448fd1e05ced4, and default selection remains forbidden. The chained decision is recorded in pancha_pakshi_1879_solar_proportional_materialization_2026_07_20.json, whose canonical SHA-256 is e31e0664090b9a38bdcd52b660c04998a0412eab223f4a84fe745b9e54d25383. It freezes the Stage 2C decision at SHA-256 b0698a4163a12dc6049cb30907d6e9dfebad790b35cc3661a41d77df89482976, the prior manifest at SHA-256 366f13deb4b213267b7a6e937b776cd3c3908178e11b29ba238fb3ed47f25e44, and binds the new manifest at SHA-256 6dbbf05383c7a4eb3eadebf70fdb1130ab5081ef83175bc387755ffda4db9121.

The governing object is PanchaPakshiSolarProportionalMaterialization, with immutable PanchaPakshiSolarProportionalMaterializationPolicy and PanchaPakshiSolarProportionalCell vessels. Its only admitted policy is solar_proportional_nominal_offsets_over_governing_half_tt_v1, explicitly a modern Moira composition rather than an 1879 source claim.

Stage 2A supplies the governing local-solar half and unchanged nominal schedule. Every exact nominal endpoint is retained as a reduced fraction of the full thirty-nazhigai schedule. The anchor and solar-half end are converted through the same configured reader to TT; interior endpoints are derived independently from the common anchor and complete TT span, then projected to UT1. The outer TT and UT1 endpoints are exactly the governing solar bounds. The result contains 25 positive contiguous half-open cells and performs no fixed-second conversion, clipping, wrapping, repetition, tail fabrication, or duration accumulation.

The low-level engine, facade, and REST surfaces are respectively pancha_pakshi_solar_proportional_materialization_at(...), Moira.pancha_pakshi_solar_proportional_materialization(...), and POST /v1/pancha-pakshi/schedule/solar-proportional. Paksha remains caller supplied; current-cell selection and astronomical paksha inference remain not_performed. The provenance routing status is solar_proportional_materialization_performed_paksha_caller_supplied_no_current_cell_or_inference.

The raw profile’s seasonal_scaling omission remains honest source-layer metadata: the 1879 witness does not attest the proportional rule. The Stage 2D result avoids claiming the same operation both omitted and performed by replacing that item with source_attested_solar_proportional_materialization, preserving the historical non-attestation while naming the separately performed modern composition. Earlier public results and the hashed profile are unchanged.

Validation is bounded to exact-fraction, independent-mapping, TT/UT1 closure, half-open topology, capability, immutability, facade/transport, and provenance invariants, including rejection of forged fraction/endpoint mappings and contradictory Stage 2D routing provenance. The existing JPL Horizons comparison remains authority evidence only for the inherited topocentric solar anchor. No external Pancha Pakshi proportional-timing oracle, independent-witness corroboration, current-cell claim, or new astronomical or historical accuracy claim is asserted.

The detached result validates fraction-to-TT mapping, outer UT1 closure, and UT1 ordering/contiguity from its fields. Reader-dependent interior TT-to-UT1 inverse truth is validated on the factory path, where the configured reader is available; the focused linear-clock test checks every interior endpoint rather than treating a detached vessel as an astronomical oracle.

Stage 2D Implementation Verification

The admitted implementation was verified on 2026-07-20 with the project Python 3.14.3 .venv, MOIRA_TEST_MODE=1, MOIRA_STRICT_KNOWN_ISSUES=1, and downloads disabled:

This receipt validates the implemented policy and transport contract. It does not turn DE441, the internal structural invariants, or the inherited Horizons solar-boundary fixture into an external Pancha Pakshi proportional-timing oracle.

Stage 2C Fixed-Clock Current-Cell Addendum

Stage 2C admits one further capability, fixed_clock_current_cell_selection, for the same source_scoped_public profile. The profile document remains unchanged at canonical SHA-256 876e4cc7cc5d894f5e558ac733913e84a8b779f72c77661e89d448fd1e05ced4, and default selection remains forbidden. The chained decision is recorded in pancha_pakshi_1879_fixed_clock_current_cell_2026_07_20.json, whose canonical SHA-256 is b0698a4163a12dc6049cb30907d6e9dfebad790b35cc3661a41d77df89482976. It freezes the Stage 2B decision at SHA-256 67cd0ac7cae74556dce702deb29708a5e99a4d19c79184444e3a81d903934449, the prior manifest at SHA-256 766f92650bc050f4c88670f8fd6307036ff49a97e812c9efc9a428fb76e53e17, and binds the new manifest at SHA-256 366f13deb4b213267b7a6e937b776cd3c3908178e11b29ba238fb3ed47f25e44.

The governing object is PanchaPakshiFixedClockCurrentCellSelection, with the immutable PanchaPakshiFixedClockCurrentCellSelectionPolicy and finite PanchaPakshiCurrentCellSelectionStatus values selected and unmaterialized_solar_half_tail. Its only admitted policy is fixed_clock_current_cell_half_open_solar_precedence_v1. This is explicitly a modern Moira interval-membership policy, not an additional rule attributed to the 1879 witness.

The policy resolves the governing half-open local-solar half before selecting from that half’s Stage 2B materialization. The requested instant is converted once to reader-bound TT and belongs to a cell exactly when start_jd_tt <= requested_jd_tt < end_jd_tt. Membership tolerance is 0.0 s: shared endpoints belong to the following cell, the fixed end is excluded, and the Stage 2B 0.0001 s topology coalescence cannot alter current-cell ownership. At exact sunset or sunrise, the newly governing solar half wins; post-boundary cells from the previous short half remain inspectable but cannot remain current.

When a long solar half outlasts its fixed span, the result reports unmaterialized_solar_half_tail and current_cell=None. It never clips, wraps, repeats, stretches, borrows, or retains a cell. Paksha remains caller supplied; astronomical paksha inference and solar-proportional scaling remain not_performed.

The low-level engine, facade, and REST surfaces are respectively pancha_pakshi_fixed_clock_current_cell_at(...), Moira.pancha_pakshi_fixed_clock_current_cell(...), and POST /v1/pancha-pakshi/schedule/fixed-clock/current-cell. The provenance routing status is fixed_clock_current_cell_selection_performed_paksha_caller_supplied_no_scaling_or_inference. Validation is limited to structural and physical invariants over the already admitted Stage 2B cells: midpoint and boundary membership, solar-half precedence, explicit tail behavior, capability gating, immutable result consistency, and strict facade/transport policy. No external current-cell oracle, independent-witness corroboration, or new astronomical or historical accuracy claim is asserted.

Stage 2B Fixed-Clock Materialization Addendum

Stage 2B admits one further capability, fixed_clock_materialization, for the same source_scoped_public profile. The profile document remains unchanged at canonical SHA-256 876e4cc7cc5d894f5e558ac733913e84a8b779f72c77661e89d448fd1e05ced4, and default selection remains forbidden. The chained decision is recorded in pancha_pakshi_1879_fixed_clock_materialization_2026_07_20.json. It freezes the prior Stage 2A decision at SHA-256 de8e40c161a327695702b9b152f89da8e848f32aafb4d0b155176d28381c9fd2, the prior manifest at SHA-256 4587306ded9b5760940e7f80c45b6c40132590473e910ea9350c9d7fa141a2ee, and binds the new manifest at SHA-256 766f92650bc050f4c88670f8fd6307036ff49a97e812c9efc9a428fb76e53e17.

The governing object is PanchaPakshiFixedClockMaterialization, with immutable PanchaPakshiFixedClockMaterializationPolicy and PanchaPakshiFixedClockCell vessels. Its only admitted policy is fixed_24_minute_nazhigai_from_local_solar_half_start_v1, explicitly a modern Moira composition rather than an 1879 source claim. The assembly is bounded:

The source roles remain separate. The 1879 leaves named in the decision fixture govern the nominal schedule only. The University of Madras Tamil Lexicon, page 2231 defines a nazhigai as sixty vinadi or twenty-four minutes. The IERS TT convention and IERS Technical Note 29 govern TT’s conventional realization and SI-second basis. The Stage 2A Horizons comparison governs only the local-solar anchor. No one source is presented as attesting the full modern composition.

The low-level engine, facade, and REST surfaces are respectively pancha_pakshi_fixed_clock_materialization_at(...), Moira.pancha_pakshi_fixed_clock_materialization(...), and POST /v1/pancha-pakshi/schedule/fixed-clock. The provenance routing status is fixed_clock_materialization_performed_paksha_caller_supplied_no_current_cell. The result exposes every materialized cell and the fixed-versus-solar boundary topology, but deliberately contains no current-cell selection and performs no solar-proportional scaling.

Stage 2A Local-Solar Context Addendum

Stage 2A admits one additional capability, astronomical_context, for the existing agastya_madras_1879_akshara_fixed_clock profile. Admission remains source_scoped_public, default selection remains forbidden, and the profile document is unchanged at canonical SHA-256 876e4cc7cc5d894f5e558ac733913e84a8b779f72c77661e89d448fd1e05ced4. The additive decision is recorded in pancha_pakshi_1879_local_solar_context_2026_07_20.json. It hash-binds the frozen Phase 1 decision and the manifest transition rather than modifying the historical fixture.

The admitted governing object is PanchaPakshiLocalSolarContext with its embedded immutable PanchaPakshiLocalSolarContextPolicy. The policy ID is local_solar_day_explicit_paksha_v1. It is explicitly a modern Moira composition policy, not a source-attested reading from the 1879 print.

The low-level engine accepts a UT1 instant, location, explicit profile, and a caller-supplied Purva or Amara source label. Datetime-facing facade and REST surfaces accept a timezone-aware datetime, normalize it to UTC, and preserve the established UTC civil-anchor conversion into UT1. The existing local-solar-day boundary then:

The result exposes requested and solar-event UT1 JDs, location, derived half and weekday, explicit paksha, policy, nominal schedule, and source-scoped provenance. Its routing status is local_solar_half_and_weekday_performed_paksha_caller_supplied. Polar geometry without lawful bounds fails explicitly.

Stage 2A does not infer paksha from the Moon, accept natal identity, scale nominal durations, materialize nazhigai offsets as instants, claim a current activity, add subdivisions or authority birds, score conditions, search windows, collate witnesses, or raise the profile to corroborated_public. Those boundaries remain separate work.

The Phase 1 computational-semantics digest remains unchanged because it covers the source profile’s identity, nominal schedule, durations, and relationships; it does not validate the additive context policy. Stage 2A evidence instead consists of policy identity, profile/manifest admission chaining, established local-solar boundary tests, context assembly invariants, facade/REST UTC normalization, and explicit failure behavior. No external Pancha Pakshi oracle is claimed for this modern composition.

The astronomical boundary itself has a narrower primary-authority check. The offline JPL Horizons sun-new-york-equinox observer-table fixture permits 2 s; the content-identified DE-0441LE-0441 validation run differed by 0.082 s at sunrise and 0.123 s at sunset. The Stage 2A decision fixture hash-binds that authority fixture and records the comparison semantics. This supports only the topocentric local-solar boundary, not the historical schedule or any broader Pancha Pakshi doctrine.

The comparison is not presented as identical-threshold parity: Moira uses its declared -0.833-degree crossing, while the authority fixture was derived at -0.8333 degrees and labels time as UT. Both differences are explicit inside the 2 s gate.

Source-Scoped Public Admission Addendum

Phase 0 replaced the former private-or-universal admission binary with finite profile states: research_only, source_scoped_public, and corroborated_public. No state authorizes a default; manifest schema 2 rejects default_selection_allowed=true.

Phase 1 admits agastya_madras_1879_akshara_fixed_clock as source_scoped_public. The exact public claim is the Agastya-attributed Madras 1879 aksara/query-or-name-initial fixed-clock Pancha Pakshi operating schedule and directed relationship matrix. Its admitted capabilities are aksara_identity, nominal_schedule, and directed_relationships.

This decision does not convert the machine reading into independent-witness consensus or a linguistic-authority claim. Those limits remain explicit and prevent a corroborated, generalized, natal, or default-canon claim. They do not prevent publication of this narrowly named witness product with its provenance and omissions attached to every result.

The additive public admission fixture preserves the former profile SHA-256 02f1252cbcff10f680148b0213021d30db043c0ecc7387be727ad5d60de04e98, binds schema-v2 profile SHA-256 876e4cc7cc5d894f5e558ac733913e84a8b779f72c77661e89d448fd1e05ced4, and records the unchanged computational projection across 10 identity symbols, 28 schedules, 700 cells, and 20 directed relationships. Admission and provenance metadata are excluded from that projection because this migration intentionally changes them. The original blind, grid-reading, adjudication, and reconciliation records remain frozen in their historical private state.

At the Phase 1 decision, public access was additive through moira.pancha_pakshi, package-root and moira.vedic exports, five kernel-free Moira methods, and five explicit-profile REST routes under /v1/pancha-pakshi. There was no native path or astronomical routing in that Phase 1 product. The Stage 2A addendum above preserves those operations and adds one separately governed kernel-backed context operation.

The non-bundling policy is unchanged: Moira distributes no archival scan, derivative PDF, OCR, page image, copied layout, source prose, or third-party translation. That architectural boundary is not a rights-clearance phase.

Historical Machine-Reconciliation Decision

The remainder of this document records the state immediately before the source-scoped public decision above. Statements that the profile was private or that transport was deferred are retained as historical evidence and are superseded by this addendum and the live admission standard.

At that historical checkpoint, the agastya_madras_1879_akshara_fixed_clock profile was private and research_only. A multi-pass page-image adjudication resolved the six material disagreements preserved by the original blind and representative-grid records. It established the complete 30/5/6 day/night scope, paired-weekday semantics, and table axes; confirmed the Amara-night schedule and all twenty directed nonself relationship cells; and proved that the existing Pūrva-night generator was wrong. The profile record then acquired the corrected Pūrva-night step and offsets.

This was machine-assisted resolution for the then-private profile, not a linguistic-authority claim, independent-witness consensus, or public doctrine. That historical reconciliation itself added no package-root export, Moira facade method, FastAPI route, OpenAPI schema, or native path; the later Phase 1, Stage 2A, and Stage 2B decisions above admitted the bounded public surfaces separately.

The machine-readable evidence consists of four distinct records:

The versioned reconciliation remains marked private_executable_machine_reconciled because it records that historical checkpoint. It binds the corrected profile and preserves the earlier records without rewriting them. That record is not an external oracle and does not by itself authorize admission; the later additive decisions do.

Governing Witness And Method

The source remains the 1879 Madras print at dli.rmrl.000451_images. The locally inspected Internet Archive derivative PDF has SHA-256 ed52945ee141faa3f6967b8f043077b95abef9ff674ffb83eaba633417c669c9. Internet Archive metadata distinguishes that derivative PDF from the original image ZIP; both file identities and archive hashes are retained in the profile provenance.

Two initial read-only reviews were kept separate and frozen before the first reconciliation:

  1. a blind extraction that did not inspect Moira’s profile, JSON, tests, or doctrine documents before freezing its reading; and
  2. a separate visual reading of representative printed schedule grids.

The later adjudication split temporal, fraction, schedule/table-axis, and relationship questions into independent page-image tasks. Blind tasks did not inspect the profile before freezing their readings. Each record names its reviewer identity, protocol, witness hashes, timestamp, and per-reading leaves. The reconciliation binds all three record hashes and the corrected profile hash. OCR was used only to locate leaves; page images governed every recorded reading. This is independently produced machine-assisted evidence with its uncertainty visible and no human-review dependency.

Machine-Adjudicated Readings

The follow-up page-image adjudication records these normalized facts for the then-private, now source-scoped-public profile:

These readings are asserted against the source-scoped profile by focused data-integrity tests. They remain machine-read source evidence, not a claim that the profile is a universal Pancha Pakshi canon.

Resolved Machine Findings

Finding Historical machine-reconciliation result Remaining boundary
Temporal-model scope IA leaves n6 and n15 establish the fixed 30/5/6 structure for both day and night. Independent-witness collation remains necessary for a corroborated claim; no human-review dependency applies to the bounded source-scoped product.
Paired headings and table axes Paired names are weekday alternatives. Vākkiya grids select weekday assignments; Eḻuttu and Toḻil grids expose samam/activity/bird assignments under their identified axes. Visual grid order is never chronological authority.
Pūrva-night assembly The prior generator was wrong. Sunday/Tuesday begins Crow-Eat, Owl-Rule, Vulture-Die, Peacock-Walk, Cock-Sleep; step and offsets are corrected in the profile. The corrected 175-cell surface is public only as part of the named source-bound product.
Amara-night seeds and assembly Sun–Sat seeds remain Vulture, Cock, Vulture, Owl, Crow, Peacock, Cock; the existing step, offsets, chronology, and assignments are confirmed. No universal doctrine is inferred.
Directed relationships IA leaf n52 directly defines all twenty ordered nonself cells; direction is subject-to-target and reciprocity is never inferred. Independent-witness collation remains incomplete.
Activity synonyms Alternate Eat and Sleep lexemes are computational synonyms within the same five-state scheme. No additional state is manufactured from lexical variation.

The earlier disagreements were substantive and remain visible in their frozen records. The later adjudication resolves them by identifying the governing axes, grammar, and text layer; it does not pretend that the earlier readings always agreed.

Source Artifact Boundary

The 1879 print and all later witnesses are research references, never package assets. Moira distributes its own code, schema, prose, and independently normalized symbolic profile; archive files and copied source expression remain outside the product.

The standing boundary is therefore architectural rather than a clearance exercise:

Reproducible Tamil Source-Reading Packet

This packet preserves the exact hashed witness, leaf locators, and machine readings so that any future source review can reproduce, confirm, or reject them. It creates no human-language-review dependency:

  1. n6, n10, n18, and n50–n51 assign 1 1/4, 1 1/2, 2, 3/4, 1/2 nazhigai to Eat, Walk, Rule, Sleep, and Die respectively.
  2. Paired weekday headings denote two discrete alternatives sharing one complete table.
  3. The identified table axes assign birds and activities; explicit prose and verse, not visual grid order, govern chronology.
  4. Pūrva-night uses step 1, offsets Eat 0, Walk 2, Rule -1, Sleep 1, Die -2, and chronology Eat, Rule, Die, Walk, Sleep.
  5. Wednesday, Thursday, and Saturday Amara-night first-Eat birds are Owl, Crow, and Cock.
  6. IA leaf n52 defines every ordered nonself relationship directly under subject-to-target grammar.
  7. Alternate Eat and Sleep terms are lexical synonyms within the same five computational activities.

Priority visual leaves are n5–n6, n10, n16–n18, n21–n25, n26–n35, n41–n42, n50–n52. The archive PDF page P corresponds to IA leaf n(P-1) in the inspected file.

Historical Admission Gate Ledger

This table records the pre-admission checkpoint; the addenda above state the current bounded public status.

Gate State at the historical checkpoint
Machine-assisted transcription and adjudication Completed for the then-private profile; all six prior findings resolved without rewriting the frozen records.
Human-language review Not an admission dependency; no reviewer is required.
Independent-witness collation Not completed; the 1867 parallel relationship reading is provisional only.
Assignment/chronology precedence Resolved for the named source profile: identified grids govern assignments; explicit prose and verse govern chronology.
Source-artifact policy Satisfied: witnesses are reference-only and never bundled; this is not an admission gate.
Identity product Satisfied for the named aksara/query-or-name-initial product; explicitly not natal Moon identity.
Source-owned examples Resolved into a complete source-owned Pūrva-night assignment oracle and confirmed Amara-night examples.
Public vessels and transport Deferred at this checkpoint; admitted later by the Phase 1 and Stage 2A through Stage 2I decisions above.

Validation Receipt

The corrected checkpoint was verified on 2026-07-20 with the project Python 3.14.3 .venv, MOIRA_TEST_MODE=1, MOIRA_STRICT_KNOWN_ISSUES=1, and downloads disabled:

This historical receipt proves the then-private schema/hash integrity, fail-closed loading, the corrected source-owned Pūrva-night assignments, generic schedule invariants, and preservation of the shared Python solar-day boundary. The two real solar-window tests used the discovered DE441 resource; no Pancha Pakshi native path exists. The adjudication remains machine-assisted research evidence, so no linguistic-authority claim, external-oracle parity, universal-doctrine claim, or admission beyond the later explicitly bounded decisions is asserted.

Constitutional Eight-Phase Closure And Later SCP Continuation

  1. Constitutional Phase 2 is closed at the private research boundary. Its exact corpus comprises five EAT, five WALK, five RULE, four SLEEP, and five DIE classifications, with verse 250 retained solely as an unclassifiable text-layer identity conflict. Do not add another classification unless its Phase 1 truth, source locator, uncertainty, and derivation rule are first preserved.
  2. Constitutional Phase 3 is closed at the private research boundary. Its derived convenience views, exact absence behavior, and vessel hardening add no source semantics, interpretation, doctrine, condition, score, prognosis, advice, selector binding, runtime profile, export, facade method, or REST route.
  3. Constitutional Phase 4 is closed at the private research boundary. It admits exactly one explicit activity-plus-ordinal lookup policy with no default, temporal inference, Stage 2K selector binding, outcome interpretation, or public surface. The two Stage 2K selectors remain separate, named, unbound candidates and neither is attributed to Uromarisi.
  4. Constitutional Phase 5 is closed at the private research boundary. Its 24 typed records preserve exact relation-clause presence, absence, bounded surface categories, confidence, and source bindings without inferring endpoints, direction, meaning, favorability, condition, or score.
  5. Constitutional Phase 6 is closed at the private research boundary. Its derived views explicitly distinguish detected, admitted, and scored subsets, harden typed membership, and preserve exact lookup absence without resolving relation meaning or widening runtime admission.
  6. Constitutional Phase 7 is closed at the private research boundary. Its 24 policy-bound profiles integrate exact classification and relation truth while remaining explicitly not evaluable, unscored, and non-prognostic.
  7. Constitutional Phase 8 is closed at the private research boundary. Its immutable aggregate records complete structural partitions without ranking, weighting, judging, scoring, or forecasting. The agreed Phases 1–8 are complete. That boundary required a new explicit Phase 9 decision, which the later SCP continuation below supplies without retroactively changing Phase 8.
  8. Stage 2M exactly identifies the Commissionerate A5 Ramadevar Panchapakshi candidate and excludes several Patchani false cognates, but the target remains title-only. Actual A5 leaves or a critical transcription plus item-level physical, custodial, and copying-history metadata are needed only before a corroborated_public claim and do not alter private closure.
  9. Preserve every disagreement by witness and text layer; do not merge doctrines or repair them by symmetry. The verse 256+ illness-duration material is a separate prospective product, not unfinished coverage in this corpus.
  10. Preserve existing admitted products at their current boundaries: Stage 2D/E materialization and current-cell selection remain modern; Stage 2F paksha inference remains separate; Stage 2H Padu remains a pure Paksha-by-weekday lookup; and Stage 2I first-EAT remains a pure first-samam seed lookup. None silently acquires Uromarisi outcome, authority, condition, score, electional, or forecasting semantics.

SCP Phases 9-12 Closeout

The explicit continuation decision authorizes SCP Phases 9-12 but adds no new historical source truth. Phase 9 creates 24 canonical private nodes and 17 detected-clause candidate annotations. Endpoints and direction remain unestablished, so admitted and scored edge sets are empty and graph metrics are not evaluable. Its decision SHA-256 is 49935df6e96595b5cd365dbea12acabcf862eb81a120c3d9122d29ad4962872b.

Phase 10 freezes the complete cross-layer chain under structural SHA-256 2133ad1c72ea5209facbb83ff8f40cfd09c1efea5340df7943fe08ff599cface. Its focused adversarial contract rejects ordering, identity, binding, conflict, policy, edge, scoring, and metric substitutions. The decision SHA-256 is 9ef977585ad1dc9dc517316eb864a8de26f462fb852977bfef936d8756ef64a0.

Phase 11 freezes the backend architecture and validation codex only after Phase 10 establishes executable invariants. Its decision SHA-256 is 697eecaf22cf4e8d42ca9b7044633e6407ca8d5577dd4407180029cfc00055c0.

Phase 12 admits only the immutable constitutional-status vessel and package/ facade operation. Its governance-only REST route returns the same status with no historical or network data and creates no manifest profile. It preserves research_only, forbidden medical use, no prognosis, and unavailable relation/graph/condition semantics. Its decision SHA-256 is 47ddf7348a235245ca89dc637a20be37b7bec244aa1ded968c9c04b884d05a6d. No human-language reviewer or Commissionerate A5 artifact is required for this closure; A5 remains relevant only to a future corroborated_public claim.

The present resting place is four tested, source-scoped public profiles: the corrected 1879 aksara schedule with the separately named Stage 2A through Stage 2E modern compositions, Stage 2F source-mapped astronomical inference, and Stage 2I source-attested first-samam EAT-seed lookup, plus the separate 2024 nakshatra table and Stage 2G modern natal-Moon composition, plus the Stage 2H 2024 Padu-only table, plus the Stage 2K exact Sookshma temporal-selector profile with two mandatory explicit policies, plus the Stage 2N explicit modern schedule composition with no clock or outcome binding, plus Stage 2O civil-time routing under mandatory timing and selector policies with no fallback or outcome binding, plus the Stage 2P research-only illness locator grid and Stage 2Q five-cell EAT semantic-atom pilot with no runtime payload or medical-truth claim, plus the Stage 2R five-cell WALK extension, Stage 2S five-cell RULE extension, Stage 2T four-cell SLEEP extension with verse 250 blocked under the same boundary, Stage 2U five-cell DIE extension, and the private constitutional Phase 2 classification closure over all 24 unconflicted records, and the private constitutional Phase 3 derived inspectability and vessel-hardening closure, plus the private constitutional Phase 4 explicit activity-plus-ordinal policy closure with both temporal selectors left unbound, plus the private constitutional Phase 5 relation corpus over all 24 classified cells with no endpoint, direction, or scoring meaning, plus the private constitutional Phase 6 hardening and inspectability closure with 17 detected, 0 admitted, and 0 scored relations, plus the private constitutional Phase 7 corpus of 24 structurally integrated but explicitly not-evaluable local conditions, plus the private constitutional Phase 8 structural aggregate, the private Phase 9 network projection, the private Phase 10 hardening receipt, the Phase 11 architecture freeze and validation codex, and the Phase 12 governance-status-only public surface. None is a universal canon, and unresolved or alternate doctrine must not be silently inferred through the public API.

This site is open source. Improve this page.