[BernardLowe]

Add optional Go/CUDA KZG opening computation without Python or CGO. Core retains the canonical scalar mapping, commitment/proof formats, and root verification; the default CPU and portable builds do not depend on CUDA.

The independent accelerators/cuda module loads CUDA Driver/NVRTC, checks pinned ICICLE arithmetic headers, and reuses one context, fixed SRS table and bounded batched/stream workspaces. It includes Windows Git Bash preparation/smoke scripts and a no-device CI job for Linux/Windows. Gateway service admission remains outside Core.

Validation:

• Core standalone full test/vet/build passed; targeted KZG regression tests passed.
• Optional module standalone race tests, vet, CGO-disabled Linux build and Windows cross-build passed.
• Actual Go FFI NVRTC 12.9 compilation of the final kernel to compute_86 PTX passed without a GPU.
• New independent reviews passed after fixing typed-nil rejection and pinned, same-stream HtoD ordering.
• Hardware execution and performance are pending RTX 3080 smoke; this host has no NVIDIA GPU. Historical evaluator GPU results do not qualify this implementation.
• Selected Core/Gateway/runtime/evaluator boundary checker passed, including explicit Gateway adapters, migration audit (292 checks) and all 15 capability parity vectors. Integration pins and releases are unchanged.
• Linux and Windows optional-module CI and the Core CI passed at this head.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-20T07:29:28.025928Z	b37f987	Draft marked ready

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.