EIP-712 multisig review fixes: tests and docs

[WiktorStarczewski]

Review fixes on top of #3856: tests that now fail when the behaviour they name breaks, and MASM docs that match the code. Only tests, comments and one local constant name change; no MASM instruction changes.

• Rejection tests assert the ECDSA verifier's own failure, so they fail if the EIP-712 branch is disabled
• New test pins raw-signature precedence over a malformed EIP-712 entry, for ECDSA and Falcon approvers
• Every fee cycle-bound guard (regular, guarded and smart multisig) also measures an EIP-712 signer
• The generic EIP-712 vector test uses distinct input bytes, so limb and byte order are pinned
• Docs and names: the verify_bytes stack comment, PUBLIC_KEY_COMMITMENT_LOC, the auth_tx advice-map note, the guardian's EIP-712 option, the preimage packing comment

One commit per fix, based on 0xMiden#3856 at 54b5cfac5. Please merge this into the 0xMiden#3856 branch before 0xMiden#3856 lands: deleting that branch would close this PR.

Verification

• Gates on the tip: make format-check, typos-check, toml-check, shear, clippy, clippy-no-std, doc, test-docs and test-release (1,875 tests), all green.
• Each test change has a mutation proof. With the EIP-712 branch forced off, the rewritten rejection cases, the double-count control and the three new fee-guard cases fail. With the raw branch forced off, both precedence cases fail on the EIP-712 checks. With the generic test's limb push order reversed, the old symmetric inputs pass and the new inputs fail. Every case passes on the real code.

Commits

• d43e6a10c fix(rev): round 1 - C-02 F-002 EIP-712 rejection tests prove the branch ran
• 4b13cf714 fix(rev): round 1 - C-03 F-005 verify_bytes stack comment order
• 7b1d74b6d fix(rev): round 1 - C-03 F-006 name the adapter local after PK_COMM
• 997e112b0 fix(rev): round 1 - C-03 F-007 both signature formats come from the advice map
• cde91f4dd fix(rev): round 1 - C-03 F-008 guardian doc names the EIP-712 option
• 74244c978 fix(rev): round 2 - C-06 F-009 pin raw-signature precedence over EIP-712 entries
• 851edc91a fix(rev): round 3 - C-03 F-012 preimage packing comment sits on its loop
• 1ec00dffb fix(rev): round 4 - C-08 F-014 cycle-bound guards measure an EIP-712 signer
• 805957f63 fix(rev): round 4 - C-09 F-015 generic EIP-712 test uses asymmetric inputs

Reviewers: 1ec00df is the largest change - it threads one SignatureFormat helper through the three fee-payment test harnesses; everything else is local to one test or one comment.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Advanced

Run ID: 9a0e1a5d-84af-4121-9512-27e1bc69ebc0

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}