generator: 4 new speculations across co-emergence and gaussian-gravitational-decoherence

[Copilot]

Executes the generator routine's daily cycle: read the claim graph, mine dead-claim tombstones and live-claim gaps for falsifiable physics speculations, and file them per tools/claim_graph.md schema.

Speculations added

• ce-second-iterate-radius-power-law — a specific decay rate (rho_2(beta) ~ C/beta) for the second-iterate spectral radius, filling in a mechanism ce-second-iterate-real-spectrum flags as unexplained. Falsifiable by fitting the exponent across beta and N.
• ce-ftoy-attractive-coupling-subcritical — asks whether the untested gamma > 0 (attractive) branch of F_toy loses stability via a subcritical/hysteretic bifurcation, distinct from the supercritical one established on the repulsive branch and from the already-killed fixed-point-multiplicity question.
• ce-mass-phase-residual-in-ggd-kernel — cross-program: chains co-emergence's mass↔signature identification through to gaussian-gravitational-decoherence's noise kernel, asking whether a phase residual survives that ggd-noise-kernel's real bilinear form omits. Flags an axiom-smuggling risk in comparing a timeless fixed-point phase to a Schrödinger-picture instant.
• ggd-cpmg-odd-even-colored-asymmetry — mined from ggd-cpmg-degeneracy's tombstone: tests whether the odd/even CPMG filter-function asymmetry found there survives against a colored (non-static) noise spectrum, independent of the open phonon-mechanism question.

Validation

• Cross-checked all four against OBJECTIVES.md in both affected programs to avoid restating any existing named milestone.
• Cross-checked against existing claims (live and dead) for near-duplicates; declared depends_on context edges to the nearest neighbors with an explicit statement of how each new claim differs.
• python tools/claim_graph.py lint introduces no new errors or warnings.
• Left signature-change-boundary untouched (thread halted under needs-human).