Global meridian-index ATTACK packet

Research note, 8 September 2026. Not a new Publication.

The packet proves conditional nonvanishing of the actual generic native six-root meridian cokernel using a global tetrahedron-index functional. Hypotheses: exact actual moment-character consistency, nonnegative filling angles, and no zero-degree ray in the homogeneous nonnegative cone. The functional detects the original algebraic cokernel unit. It does not yet detect the filled-vacuum class $1-l^2$, prove unrestricted filling nonvanishing, or identify a surgery invariant.

Read FPRD-global-meridian-index-detector-2026-09-08.md first, then hostile-audit.md and source-ledger.md. The proposed FPRD-research-loop-state.md is the handoff for the next ATTACK. Prior native peripheral/collar and intrinsic-filling comparisons retain all their caveats.

Reproduce the exact checks

From this directory, run:

python3 verify-global-meridian-index.py

Python 3.10 or newer; standard library only. Expected result: 24 exact checks pass. verification-output.txt records the run, and exact-results.json separates computational checks from mathematical conclusions. The proof of global convergence and nonvanishing is in the note, not inferred from seed-series truncation.

The checks manipulate general skew forms, polynomial degree identities and a local seed normalization. They do not enumerate manifolds, slopes, triangulations, sign assignments or skein configurations. The one-block weak-angle distinction is an abstract lattice model, not an actual triangulation.

Integrity and scope

SHA256SUMS.txt covers the nine primary payload files, including this README, but not itself or the archive. global-meridian-index-attack.zip contains those payload files and their manifest. ARCHIVE-SHA256.txt separately records the archive hash. Primary source texts are not redistributed; exact inspected versions and retained-copy hashes are listed in the source ledger and metadata.

The next mode is ATTACK on the longitude difference or on a new detector at the null-ray boundary. AI-assisted, internally audited work; no external specialist review, author endorsement or priority claim.