docs: update verify-stark docs + rename verify_proof_output

[stephenh-axiom-xyz]

Updates verify-stark guest library docs do explain DEF_IDX more. Also renames verify_proof_output to verify_stark

[github-actions]

Claude finished @stephenh-axiom-xyz's task in 45s —— View job

---

Code Review: PR #2600

• Gather context and read changed files
• Review changes
• Post review feedback

---

Overall: Looks good! Clean rename and helpful docs expansion. A couple of minor items:

Typo

docs/vocs/docs/pages/book/guest-libraries/verify-stark.mdx:30 — "Cals" should be "Calls":

-Cals `get_proof_output` and panics if the result does not match `expected`.
+Calls `get_proof_output` and panics if the result does not match `expected`.

Fix this →

Rename

The verify_proof_output → verify_stark rename is complete — no stale references remain in the codebase.

Docs

The new DeferralExtension section is a solid addition. The explanation of DEF_IDX indexing and the SDK wiring example are clear and practical.

One minor suggestion: the code example references several types (VerifyProver, VerifyCircuitProver, DeferralProver, DeferralFn, verify_stark_deferral_fn, etc.) without import paths. Consider adding a brief use block or a comment indicating which crates these come from, so readers can locate them more easily.

[github-actions]

Claude finished @stephenh-axiom-xyz's task in 50s —— View job

---

Code Review: PR #2600

• Gather context and read changed files
• Review changes and provide feedback

---

LGTM — clean, straightforward PR. A few observations:

Rename (verify_proof_output → verify_stark)

• No remaining references to the old name anywhere in the repo — the rename is complete.
• The new name is more concise and consistent with the crate/library naming (verify-stark).

Docs improvements

• The expanded DEF_IDX explanation and the new "Using DeferralExtension" section are a solid addition. The link from the DEF_IDX description to the #using-deferralextension anchor ties things together well.
• The code example references (VerifyCircuitProver, DeferralProver, verify_stark_deferral_fn, make_extension) all correspond to real types/functions in the codebase.

Minor nit (non-blocking):

• docs/vocs/docs/pages/book/guest-libraries/verify-stark.mdx:5 — the source code link points to main branch (https://github.com/openvm-org/openvm/tree/main/guest-libs/verify-stark). If develop-v2.0.0-beta diverges significantly from main, this link could be confusing for readers. Not introduced by this PR though, so fine to leave as-is.

No issues found.

[github-actions]

Claude finished @stephenh-axiom-xyz's task in 1m 2s —— View job

---

Code Review: PR #2600

• Gather context and read changed files
• Review changes against base branch
• Post review feedback

---

LGTM — clean rename and solid docs expansion. All three changed files are consistent and no stale references to the old names remain.

Rename summary:

• get_proof_output → verify_stark_unchecked (lower-level, returns ProofOutput)
• verify_proof_output → verify_stark (convenience wrapper that panics on mismatch)
• Names are now consistent with the crate/library naming (verify-stark) and more descriptive

Docs:

• The expanded DEF_IDX explanation and new "Using DeferralExtension" section are clear and practical
• Good ordering: verify_stark_unchecked (lower-level) is described first, then verify_stark as a convenience wrapper — this addresses jonathanpwang's earlier feedback about explaining verify_stark relative to get_proof_output

Re: jonathanpwang's review comments:

1. "describe get_proof_output after verify_stark" — The latest commit renamed it to verify_stark_unchecked, which makes the relationship clearer. The current ordering (unchecked first, then checked wrapper) reads naturally since the wrapper is defined in terms of the unchecked version. Either ordering works, but I'd defer to the reviewer's preference here.
2. "call using DEF_IDX" — Looking at verify-stark.mdx:26, the example verify_stark::<0> doesn't have a comment explaining what 0 is. Consider adding a brief comment in the example program:

   verify_stark::<0>(&input_commit, &expected); // DEF_IDX = 0, first (and only) deferral circuit

   Same applies to crates/sdk/programs/examples/verify-stark.rs:26.
3. "explain with comment what 0 is" — Still unaddressed in the code. The 0 in verify_stark::<0> in the example program could use a comment.

One minor nit (non-blocking):

• verify-stark.mdx:62 — the code example initializes let def_idx = 0; which is great, but def_idx is only used once (passed to DeferredVerifyCpuProver::new). The connection between this def_idx and the guest-side DEF_IDX generic const could be made more explicit with a comment.

No bugs or issues found. The rename is complete and consistent.

[jonathanpwang]

lgtm with one more comment

[github-actions]

Claude finished @stephenh-axiom-xyz's task in 49s —— View job

---

Code Review: PR #2600

• Gather context and read changed files
• Review changes against base branch
• Post review feedback

---

LGTM — clean rename and well-written docs expansion.

Rename: get_proof_output → verify_stark_unchecked and verify_proof_output → verify_stark are complete. No stale references remain in the codebase.

Docs: The new content is solid:

• The "there exists a proof" clarification on line 23 addresses jonathanpwang's latest review comment nicely.
• The DeferralExtension section with the SDK wiring example is practical and well-structured.
• DEF_IDX indexing is clearly explained.

One minor suggestion (non-blocking):

crates/sdk/programs/examples/verify-stark.rs:26 — the 0 in verify_stark::<0> could use a brief comment for readers unfamiliar with the deferral framework:

verify_stark::<0>(&input_commit, &expected); // DEF_IDX = 0: first deferral circuit

This was also flagged in jonathanpwang's review comment ("explain with comment what 0 is"). Fix this →

No bugs or issues found.