refactor: qntx.openai 1.0 (EVM/SVM/TVM, spend_controls, httpx2)

[gitctrlx]

Summary

Rewrite x402-openai 0.3 as qntx-openai 1.0, product-parity with shipped @qntx/openai.

• PyPI qntx-openai, import qntx.openai. No x402_openai shim.
• Per-chain keys: evm= / svm= / tvm=. Wallets, BIP-39, max_amount, caller http_client= removed.
• EVM: exact + upto on eip155:*. SVM: exact only on solana:*. TVM: exact on tvm:-239 / tvm:-3 (not client tvm:*).
• Official spend_controls (omit = $1). Local TS-semantics prefer_network / prefer_scheme.
• Owned httpx2 402 adapter: copy of official x402AsyncTransport retry loop, fail-closed.
• Default base_url https://llm.qntx.org/v1.

Stacked commits (bottom → tip):

1. Foundation: rename, httpx2 adapter, EVM exact+upto
2. SVM exact
3. TVM exact + close() dispose
4. README / examples / URLs

Do not tag or publish from this PR. Tag v1.0.0 after merge, then yank x402-openai 0.2.0 / 0.2.1 / 0.3.0.

Test plan

• ruff / mypy / pytest on the merged tree (127 passed)
• CI on this PR

[Copilot]

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

[gitctrlx]

Code review

No issues found. Checked for bugs and CLAUDE.md compliance.

Audit of origin/main...a6276e3f77ee8479347af5e403b5a79aaf14902b (five independent reviewers: AGENTS.md, obvious bugs, git history, prior PR comments, in-file comment guidance). Two candidate findings scored below the 80 threshold and were discarded (_REMOVED TypeError is a hard delete of obsolete kwargs, not a compatibility shim; empty _NOT_YET is leftover no-op). Guidance used: https://github.com/qntx/openai/blob/main/Agents.md (this repo has no CLAUDE.md).