docs: reference system-contract bytecode via versioned artifacts

[RealiCZ]

Summary

• Replace the inline deployed-bytecode hex blobs on the system-contract spec pages (Oracle v1.0.0/v1.1.0/v2.0.0, KeylessDeploy v1.0.0, MegaAccessControl v1.0.0, MegaLimitControl v1.0.0, SequencerRegistry v1.0.0) with references to the canonical versioned artifacts under crates/system-contracts/artifacts/.
• Each Deployed bytecode: entry now shows a truncated prefix, a link to the artifact JSON (deployedBytecode field), and a cast keccak/jq one-liner that re-derives the code hash; the code hash stated on each page remains the normative identity of the bytes.
• This removes ~11 KB of hex that duplicated data already hash-verified at build time by crates/system-contracts/build.rs, following the OP Stack specs precedent: the code hash is the normative anchor, full bytes live outside the prose page.
• Docs only; no behavior change.

Verification

• Every replaced blob was compared byte-for-byte against its artifact's deployedBytecode before deletion, and every stated code hash equals the artifact's codeHash.
• The verification command shown on each page was executed for all seven artifacts and reproduces the stated code hash.
• All seven artifact links point to files present on main.

[github-actions]

🧬 Mutation testing

No results at target/mutants/mutants.out — nothing was mutated (e.g. no mutatable changes).

[github-actions]

🧬 Mutation testing — ✅ PASS

Nothing to test — no mutants were generated on the changed lines.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 6cd972833f

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".

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

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[claude]

✅ Clean

• Reviewed a docs-only change that replaces inline hex blobs on 5 system-contract spec pages with truncated prefixes plus links to the versioned artifact JSONs and a cast keccak/jq reproduction recipe.
• Verified against crates/system-contracts/artifacts/: all 7 artifact files exist, every truncated prefix matches its artifact deployedBytecode, and every Code hash: in the docs equals its artifact codeHash.
• Code hash remains the normative anchor, and the version-pinned filenames (Oracle-1.0.0.json, etc.) keep the main-branch links stable for frozen spec versions.
• No new actionable findings. No prior review threads to reconcile.

[flyq]

1. State the working directory for the verification command.
   jq -r .deployedBytecode crates/system-contracts/artifacts/… assumes the reader runs it from a repo checkout root; the spec pages are published on docs.megaeth.com where that context is not obvious.
   A short lead-in such as "From the repository root:" above the code block (once per page) would remove the ambiguity.
2. Consider one sentence making the hash's normative role explicit.
   The PR body says "the code hash stated on each page remains the normative identity of the bytes", but no page says this in so many words.
   A single sentence on the system-contracts overview page (e.g. "The code hash is the normative identity of a contract version; the linked artifact JSON is the canonical source of the full bytes.") would make the intent survive future edits.
   The truncated 12-byte prefixes are cosmetic either way — all seven start with the same Solidity prologue (0x6080604052…), so they carry no discriminating value and could equally be dropped.

[codspeed-hq]

Merging this PR will not alter performance

✅ 366 untouched benchmarks

---

_{Comparing cz/doc/bytecode-artifact-refs (29f0c80) with main (999c516)}

[claude]

⚠️ 1 finding

• Summary: Re-review of the second commit that publishes each system-contract version's full deployed bytecode under docs/spec/static/bytecode/, links it from the spec pages, adds docs/spec/AGENTS.md guidance, and introduces docs_bytecode_consistency to pin the static files 1:1 to the versioned artifacts.
• The refactor cleanly addresses the earlier immutability concern (bytes now live in-spec, no blob/main mutation surface) and the new test makes the invariant enforceable.
• Inline comments: 1 anchorable finding posted on a changed line.
• Body-only items: none.

Still open from earlier reviews: none (prior Codex thread is resolved and the previous Claude review posted a Clean verdict).

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 29f0c80ea2

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".

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

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[RealiCZ]

@flyq
Thanks for the review — both suggestions taken:

1. The verification lead-in on every page now reads "To verify the code hash, from the repository root:". It is attached to each code block (rather than once per page) so that each version section on the Oracle page stays self-contained for readers who deep-link to a specific version.
2. Added to the system-contracts overview (Deployment Semantics): "The code hash stated on each contract page is the normative identity of that contract version. The artifact JSON linked on each page is the canonical source of the full deployed bytecode."

On the truncated prefixes: kept them intentionally. The format follows the OP specs convention, where the prefix is likewise the shared Solidity prologue plus a full-bytecode link:

https://github.com/ethereum-optimism/specs/blob/main/specs/protocol/ecotone/derivation.md#gaspriceoracle-deployment

It serves as a visual cue that the value is bytecode, not as a discriminator.

[claude]

✅ Clean

• Reviewed the re-review scope: the third commit (9dfb3ee) reverted the static-bytecode-files approach; the PR now contains only doc-only changes across overview.md and six system-contract pages, replacing inline hex blobs with a truncated prefix, an artifact-JSON link, and a cast keccak ... deployedBytecode ... reproduction command.
• Verified the artifact JSON fields the new commands target (deployedBytecode, codeHash) exist in each referenced artifact, and the code hashes stated in the docs match the artifacts.
• No new actionable findings from me this round.

Resolved from earlier rounds:

• My prior [Minor] on the reverted crates/system-contracts/tests/docs_bytecode_consistency.rs skip logic — that file no longer exists in this PR.

Still open from earlier rounds (other reviewer, unchanged status — flagging for awareness, not re-litigating):

• High-Precision Timestamp not covered by the new overview.md claim — HPT has a HIGH_PRECISION_TIMESTAMP_ORACLE_CODE_HASH in crates/mega-evm/src/system/oracle.rs but no artifact JSON and no code-hash entry on high-precision-timestamp.md, so the new sentence about the linked artifact JSON being the canonical source of the full deployed bytecode does not quite hold for that page.