feat(host): add v4 proposal proof API adapter

[smtmfft]

Summary

• add /v4/proof/proposal as a thin adapter over the existing v3 Shasta batch proof path
• convert v4 l2_block_number_start/end ranges into legacy l2_block_numbers
• map legacy v3 proof/status responses into the raiko2-style v4 response envelope
• add --proof-api-version v4 to script/stress_shasta_proposal.py while keeping v3 as the default path

Regression

• mainnet v4 native regression completed for proposal 19716
• requested L2 seed range: 8407605,8407605
• resolved proposal range: L2 8407588..8407779, L1 inclusion 25477944, last_anchor_block_number 25477875
• request payload used /v4/proof/proposal shape with l2_block_number_start=8407588 and l2_block_number_end=8407779

Tests

• python3 script/test_stress_shasta_proposal.py
• python3 -m py_compile script/stress_shasta_proposal.py script/test_stress_shasta_proposal.py
• cargo test -p raiko-host v4_proposal
• cargo check -p raiko-host
• cargo fmt --all -- --check
• git diff --check

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 682c167f0a

ℹ️ About Codex in GitHub

Codex has been enabled to automatically 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".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[chatgpt-codex-connector]

Return the full proof payload for completed tasks

When /v4/proof/proposal receives a completed legacy Shasta proof, this mapping keeps only proof.proof and drops the rest of the Proof object. Several provers populate verifier/submission data outside that field (e.g. input, quote/receipt, uuid, and Shasta extra_data), so SGX/RISC0/SP1 or native callers can get status: "completed" with an incomplete or even null proof that is not equivalent to the v3 response.

Useful? React with 👍 / 👎.

[chatgpt-codex-connector]

Include all task discriminators in task_id

The generated task_id only uses the first proposal id, last proposal id, and proof type, but the adapter accepts arbitrary proposal lists and an optional prover. Distinct requests such as aggregate proposals [10,12] versus [10,11,12] (or the same proposal range with a different prover) enqueue different Shasta tasks while returning the same v4 task_id, making the identifier unsafe for client-side correlation or de-duplication.

Useful? React with 👍 / 👎.