Overview This record contains Sunflower Endpoint Rigidity and Kernel-Forced AASC Transfer, a manuscript developing an AASC endpoint-transfer treatment of the Erdős–Rado sunflower endpoint. The paper works on the fixed core–petal carrier for (n)-uniform families. For a family[\mathcal F\subseteq \binom{U}{n},]and each candidate core (C\subseteq U), the residual petal family is defined by[\mathcal F_C={S\setminus C:S\in\mathcal F,\ C\subseteq S}.]The sunflower endpoint is then expressed as the role-cardinality condition[\exists C\subseteq U\quad \nu(\mathcal F_C)\ge k,]where (\nu(\mathcal F_C)) is the residual matching number. Central Contribution The manuscript gives a kernel-forced AASC endpoint-transfer proof for the sunflower endpoint relative to calibrated proof data[(\mathcal C,\Complete^{\mathcal C}_{k,H_k},H_k),\qquadH_k\ge H_k^{\mathcal C}<\infty.] The proof isolates the residual no-sunflower countercase as a calibrated endpoint branch: the fixed core–petal carrier is preserved; the positive endpoint is core–petal role occupation; the negative branch is global non-occupation of every (k)-petal residual slot; bounded motif branches are preserved through a declared finite certificate language (\mathcal C); only the calibrated objective non-BMF residual separator is routed to the AASC no-independent-discriminator closeout. The result is not obtained by a random-restriction, spread-lemma, or entropy-compression improvement. It belongs to the AASC proof class: fixed-carrier endpoint transfer under kernel-forced admissibility, standing, reference, and irreversibility. Method and Proof Architecture The proof proceeds through the following components: Core–petal reduction:A (k)-sunflower exists iff some residual family (\mathcal F_C) has matching number at least (k). Kernel-first dependency order:Determinate same-carrier endpoint or counterexample status already requires the AASC kernel:[K={\mathrm{Adm},\mathrm{St},\mathrm{Ref},\mathrm{Irr}}.] Cost of kernel denial:Weakening reference, standing, admissibility, or irreversibility changes or destroys fixed endpoint status rather than producing a weaker version of the same endpoint object. Certificate-language layer:A finite certificate language (\mathcal C) records bounded motif certificates, product/factor records, endpoint-preserving injections, rank accounting, and entropy accounting. Calibration layer:The motif ceiling (H_k) must dominate the raw certified motif entropy[H_k^{\mathcal C}.]Product transversals and (C_5)-type tensor motifs are treated as lawful negative structures, not forbidden residual separators. Residual separator discharge:A calibrated residual branch[\RBEsep^{\mathcal C}_{k,H_k}(\mathcal F)]can stand only as an independent same-domain endpoint-status discriminator. Under local endpoint use, the AASC consequence layer excludes such a discriminator. Lean4 Audit Support This manuscript is accompanied by a Lean4 audit release: GitHub: https://github.com/somamaley-ux/AASC-Sunflower-Endpoint-Lean-Audit DOI: https://doi.org/10.5281/zenodo.21242337 The Lean release verifies the manuscript’s AASC endpoint-transfer proof-class spine, including: the fixed core–petal residual matching carrier; the four-role AASC kernel package; the calibrated certificate-language split; the objective non-BMF residual branch; local endpoint-use discipline; the no-independent-discriminator closeout; transfer from exact local countercase use to the bounded motif certificate branch. The Lean audit is not presented as an AASC-free first-principles formalization of the classical Erdős–Rado sunflower conjecture. Its claim is sharper and bounded: it machine-checks the typed AASC endpoint-transfer mechanism and theorem-spine audit surface used by the manuscript. Scope and Proof-Class Boundary This manuscript does not apologize for using AASC. Its proof class is not a conventional spread-lemma or random-restriction route. The relevant correctness questions are: whether the fixed core–petal endpoint carrier is correctly instantiated; whether local exact-countercase use has determinate same-carrier endpoint status; whether the kernel is forced by that non-degenerate endpoint status; whether the calibrated residual separator performs independent endpoint-status work; whether the AASC no-independent-discriminator closeout applies. An AASC-free reconstruction would require a separate finite certificate-extraction theorem producing (\BMF^{\mathcal C}_{k,H_k}) certificates directly. That is a parallel reconstruction route, not a prerequisite for the kernel-forced endpoint-transfer proof class. Record Contents This record includes: the main manuscript PDF; Overleaf/LaTeX source files; bibliography and reference metadata; Lean4 audit appendix; release references for the companion Lean repository and DOI; proof-class, calibration, and adversarial audit materials.