Registry-readiness: MPL-2.0 (REUSE), de-LLM, + correct Jones-from-PD (1089/1089 green) (#10)

* licensing-readiness: single MPL-2.0 (REUSE-compliant), de-LLM

NOT a registration-ready claim. Scope: licensing/REUSE/de-LLM only,
via standards/scripts/registry-readiness.sh.
- PMPL/MPL-2.0 contradiction -> single OSI MPL-2.0; SPDX normalised;
  REUSE.toml; LICENSE rebuilt; LICENSES/PMPL removed; reuse lint:
  COMPLIANT (REUSE 3.3); SPDX-in-content wrapped with REUSE-Ignore.
- de-LLM: 0-AI-MANIFEST + llm-warmup removed; NOTICE canonicalised.

HONEST CAVEAT: Pkg.test() is PRE-EXISTING RED (1 failed, 4 errored —
UndefVarError GaussCode; SQLite v3->v4) — unrelated to these licensing
changes. Skein.jl is NOT registration-ready until those code defects
are fixed. Matches the prior Julia-suite adversarial audit (Skein =
'salvageable', not defensible).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* fix: root-cause four pre-existing test errors (schema migration + GaussCode ambiguity)

Two distinct defects, four errored testsets in Pkg.test():

1. Legacy-DB schema migration (3 errored: v1/v2/v3 -> v4 in
   test/canonical_gauss_test.jl). Two layered bugs:

   a. SkeinDB() ran the CREATE_TABLES block (which includes
      "CREATE INDEX ... ON knots(diagram_format)") *before* the
      _migrate_v3_to_v4 step that adds the diagram_format column.
      On a fresh DB the full CREATE TABLE defines every column so it
      worked; on a pre-existing legacy DB the "CREATE TABLE IF NOT
      EXISTS knots" is a no-op and the new columns are absent, so the
      index DDL threw SQLiteException("no such column: diagram_format").
      Fix: split index DDL into CREATE_INDEXES and apply it strictly
      after migrations (create tables -> migrate -> create indexes).

   b. row_to_record parsed created_at/updated_at with
      DateTime(::String), which only accepts Julia's ISO-8601
      "yyyy-mm-ddTHH:MM:SS.s". Rows inserted via the schema's own
      "DEFAULT (datetime('now'))" (legacy/external rows) use SQLite's
      "yyyy-mm-dd HH:MM:SS" with a space and no fractional seconds,
      raising ArgumentError: Invalid DateTime string. Added
      _parse_db_timestamp to normalise both formats.

2. GaussCode UndefVarError (1 errored: "Gauss fallback remains
   available" in test/knot_theory_ext_test.jl). The file does both
    and , which export different
   GaussCode types, making the bare name ambiguous. Skein's storage
   API dispatches on Skein.GaussCode, so the reference is now
   explicitly qualified.

Pkg.test() goes from 1044 passed/1 failed/4 errored to
1057 passed/1 failed/0 errored. No tests deleted, skipped, or
weakened. The remaining failure (test/runtests.jl:1210, Jones
polynomial uniqueness) is a separate pre-existing deeper defect,
reported but deliberately not hacked.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* fix: compute knot-table Jones polynomials from planar diagrams (PD codes)

Real cause (pre-existing latent defect, never green since release):
  test/runtests.jl asserted all 15 prime-knot-table Jones polynomials
  are distinct, but only 6 were. src/knot_table.jl computed Jones via
  jones_polynomial_str(dt_to_gauss(dt)) — from a bare Gauss code with no
  planar embedding. The Kauffman bracket / Jones polynomial is a DIAGRAM
  invariant: per-state loop counts depend on the rotation system, which a
  Gauss code does not encode. So the bracket collapsed distinct knots
  (e.g. figure-eight 4_1 -> a single monomial instead of
  t^-2 - t^-1 + 1 - t + t^2). src/polynomials.jl predated the assertion,
  so this test was never green (latent, not a regression).

Fix (PD-based, mathematically correct):
  * src/polynomials.jl: add a planar-diagram Kauffman state sum —
    PDCode/PDCrossing, parse_pd, pd_writhe (geometric writhe read from
    the PD: positive crossing iff over-strand out-arc b == d+1 mod 2n),
    pd_bracket_polynomial (loops counted exactly as cycles of the
    perfect matching of arc-incidence + per-state smoothing pairings),
    jones_from_pd / jones_from_pd_str. Existing Gauss-code API untouched.
  * src/knot_table.jl: embed the canonical Knot Atlas (Rolfsen) PD codes
    for all 15 knots; compute jones_polynomial from the PD.
  * src/import_export.jl: import_knotinfo! stores the PD-based Jones for
    the through-7-crossing knots so DB records stay consistent with
    prime_knot(...).
  * test/runtests.jl: assert each table knot's Jones EQUALS the published
    Knot Atlas value (term-for-term), not merely that they are distinct.

Verification (computed vs published Knot Atlas, all term-for-term MATCH):
  0_1 1; 3_1 -t^-4+t^-3+t^-1; 4_1 t^-2-t^-1+1-t+t^2;
  5_1 -t^-7+t^-6-t^-5+t^-4+t^-2; 5_2 -t^-6+t^-5-t^-4+2t^-3-t^-2+t^-1;
  6_1..7_7 all verified against https://katlas.org/wiki/<name>.
  Engine independently calibrated against the published 3_1/4_1/5_1
  before integration. All 15 distinct => the uniqueness assertion now
  holds for the right (mathematically genuine) reason.

Full suite: Testing Skein tests passed (1089/1089).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: hyperpolymath

.claude/CLAUDE.md

@@ -1,5 +1,4 @@
 # SPDX-License-Identifier: MPL-2.0
-# (PMPL-1.0-or-later preferred; MPL-2.0 required for Julia ecosystem)
 # Skein.jl - Editor Configuration
 # https://editorconfig.org
 

.editorconfig

@@ -1,4 +1,4 @@
-# SPDX-License-Identifier: PMPL-1.0-or-later
+# SPDX-License-Identifier: MPL-2.0
 # RSR-compliant .gitattributes
 
 * text=auto eol=lf

.gitattributes

@@ -1,4 +1,4 @@
-# SPDX-License-Identifier: PMPL-1.0-or-later
+# SPDX-License-Identifier: MPL-2.0
 name: BoJ Server Build Trigger
 on:
   push:

.github/workflows/boj-build.yml

@@ -1,4 +1,4 @@
-# SPDX-License-Identifier: PMPL-1.0-or-later
+# SPDX-License-Identifier: MPL-2.0
 name: GitHub Pages
 
 on:

.github/workflows/casket-pages.yml

@@ -1,4 +1,4 @@
-# SPDX-License-Identifier: PMPL-1.0-or-later
+# SPDX-License-Identifier: MPL-2.0
 # governance.yml — single wrapper calling the shared estate governance bundle
 # in hyperpolymath/standards instead of carrying per-repo copies.
 #

.github/workflows/governance.yml

@@ -1,4 +1,4 @@
-# SPDX-License-Identifier: PMPL-1.0-or-later
+# SPDX-License-Identifier: MPL-2.0
 # Hypatia Neurosymbolic CI/CD Security Scan
 name: Hypatia Security Scan
 

.github/workflows/hypatia-scan.yml

@@ -1,4 +1,4 @@
-# SPDX-License-Identifier: PMPL-1.0-or-later
+# SPDX-License-Identifier: MPL-2.0
 # Instant Forge Sync - Triggers propagation to all forges on push/release
 name: Instant Sync
 

.github/workflows/instant-sync.yml

@@ -1,4 +1,4 @@
-# SPDX-License-Identifier: PMPL-1.0-or-later
+# SPDX-License-Identifier: MPL-2.0
 # SPDX-FileCopyrightText: 2026 Jonathan D.A. Jewell
 name: Mirror to Git Forges
 

.github/workflows/mirror.yml

@@ -1,4 +1,4 @@
-# SPDX-License-Identifier: PMPL-1.0-or-later
+# SPDX-License-Identifier: MPL-2.0
 # Prevention workflow - runs OpenSSF Scorecard and fails on low scores
 name: OpenSSF Scorecard Enforcer