\input{gradient_exclusion.tex} \input{ordinary_finite_exclusion.tex} \input{hyperbolic_suspension.tex} \input{uniform_cut_decay.tex} \input{ordinary_axis_exclusion.tex} \input{bounded_regular_reduction.tex} \input{cusp_tangent_exclusion.tex} \input{concurrent_regular_exclusion.tex} \input{repeated_line_exclusion.tex} \subsection{The geometric case split for frame extraction} For a cubic of degree exactly three, either the frame test \eqref{eq:frame} succeeds, or it does not. In the latter case, the projective curve $F_3=0$ belongs to exactly one of these types: \begin{center}\small \begin{tabular}{@{}p{0.33\textwidth}p{0.56\textwidth}@{}} \toprule Leading curve & No-frame exclusion\\ \midrule Smooth cubic & Corollary~\ref{cor:smoothleading}\\ Irreducible nodal cubic & Corollary~\ref{cor:ordinarycomplete}\\ Irreducible cuspidal cubic & Proposition~\ref{prop:cusptangentcomplete}\\ Line and smooth conic, transverse & Corollary~\ref{cor:ordinarycomplete}\\ Line and smooth conic, tangent & Proposition~\ref{prop:cusptangentcomplete}\\ Three nonconcurrent lines & Corollary~\ref{cor:ordinarycomplete}\\ Three concurrent distinct lines & Proposition~\ref{prop:concurrentcomplete}\\ Double line and a distinct line & Proposition~\ref{prop:repeatedcomplete}\\ Triple line & Proposition~\ref{prop:repeatedcomplete}\\ \bottomrule \end{tabular} \end{center} The list is exhaustive by factorization of a plane cubic over $\C$. Every row has now been proved, including its lower terms and all critical degeneracies. In the repeated-line rows the singular set is a whole projective line; the proof checks its isotropic directions rather than choosing an arbitrary point on it. \begin{proof}[Proof of Proposition~\ref{prop:frameextract}] Suppose the frame test fails for a pure polynomial of degree three. Its leading curve is one of the nine types in the table. The smooth case is excluded by the gradient criterion, and the ordinary-node cases by Corollary~\ref{cor:ordinarycomplete}. Propositions \ref{prop:cusptangentcomplete}, \ref{prop:concurrentcomplete} and \ref{prop:repeatedcomplete} force a frame in every other case. All possibilities give a contradiction. This proves frame extraction. \end{proof}