docs: 3 broken links remain across circuits, CRISP, and the E3 contract guide

[Toby1009]

Updated 2026-08-20. #1795, #1796, #1797 and #1802 fixed 56 of the 59 links. Three remain,
re-checked against main at 3c7172c8. The original audit, with the full section breakdown, is
in this issue's edit history.

1. circuits/README.md:85 — a directory that is not there

[`wrapper/README.md`](bin/recursive_aggregation/wrapper/README.md)

circuits/bin/recursive_aggregation/ holds c2ab_fold, c3_fold, c3_fold_kernel, c3ab_fold,
c4ab_fold, c6_fold, c6_fold_kernel, decryption_aggregator, dkg_aggregator, node_fold,
nodes_fold, and nodes_fold_kernel. There is no wrapper/.

The link is the visible end of a longer drift. The same file names wrapper/ eleven times,
including the directory tree at :25, the C2 description at :46-50, and three rows of the circuit
table at :79-82. Retargeting the link alone would leave the section describing a layout that does
not exist, so this one needs whoever knows what the wrappers became.

2. examples/CRISP/packages/crisp-contracts/README.md:55 — Greco

https://github.com/gnosisguild/interfold/tree/main/circuits/crates/libs/greco

404 after redirects. No greco directory exists anywhere in the tree. The gnosisguild half is not
the cause: that organisation name 301s to theinterfold.

3. docs/pages/write-e3-contract.mdx:51 — a pinned commit with a later path

https://github.com/theinterfold/interfold/blob/cf50fd6c10f0df8f03eecfd19cedc17b051e72a6/packages/interfold-contracts/contracts/test/MockE3Program.sol#L9

404. The commit is reachable and is an ancestor of main, but at that revision the file was at
     packages/evm/contracts/test/MockE3Program.sol. chore: rename enclave to interfold [skip-line-limit] #1592 renamed packages/evm to
     packages/interfold-contracts, and the link took the new path while keeping the old pin. Either
     half resolves on its own:

URL	Status
blob/cf50fd6c…/packages/evm/contracts/test/MockE3Program.sol	200
blob/main/packages/interfold-contracts/contracts/test/MockE3Program.sol	200

Happy to open a PR for 2 and 3 if it is useful.

[Toby1009]

PRs opened for the mechanical parts:

• fix(cargo): align crate repository metadata with workspace URL #1795 — fix(cargo): align crate repository metadata with workspace URL (section 1, 35 manifests)
• fix(docs): repair dead external links #1796 — fix(docs): repair dead external links (sections 2, 4, 5, 8)
• fix(docs): repair broken in-repo links and anchors #1797 — fix(docs): repair broken in-repo links and anchors (sections 6, 7, minus one entry — see below)

Two corrections to the original report:

circuits/README.md:85 belongs with section 3, not section 6. I filed it as a plain relative-path
break, but it is not. The link targets bin/recursive_aggregation/wrapper/README.md, and there is no
wrapper/ subtree anywhere under circuits/ — nor a fold package — even though the table directly
above the link documents wrapper/dkg/share_computation, wrapper/dkg/*, and wrapper/threshold/*
as if they exist. So the section is stale rather than the link mistyped, and fixing the path alone
would paper over it. Left untouched in #1797 and moved to the "needs a maintainer decision" bucket
alongside section 3.

CONTRIBUTING.md:8 needed a judgment call. The guide's intro promises a Changelog section that was
never written; the changelog is actually covered under Releases and Merging. #1797 makes the word plain
text as the smallest honest fix, but pointing it at #releases or writing the section are both
reasonable alternatives — happy to change it to whichever you prefer.

Section 3's six links still need target decisions before I can open that PR.

[Toby1009]

Update. Sections 1, 2, 4, 5, 6, 7, and 8 are merged (#1795, #1796, #1797). Section 3 is down to two
entries, and both turned out to be better understood than the original report claimed.

#1802 handles three of section 3. One correction to the report while doing so: I filed
crates/data/README.md:175 as a refactor casualty, but no revision in this repository has ever
contained crates/aggregator/src/repositories.rs. That link has never resolved. The file the prose
describes is crates/aggregator/src/repo.rs.

A fourth entry, docs/pages/internals/dkg.mdx:426, is gone — #1793 removed the paragraph containing it.

That leaves two, and git rename history answers both more precisely than my report did:

examples/CRISP/packages/crisp-contracts/README.md:55 — Greco. I listed this as needing domain
knowledge because no greco path survives. It is traceable: e29a5a13 renamed the Greco core
implementation at 94% similarity to circuits/lib/src/core/threshold/user_data_encryption.nr, and
8f7f0025 then split that into user_data_encryption_ct0.nr and user_data_encryption_ct1.nr.

The remaining judgment is which of several successors a prose link should point at, since one link
now has a family behind it — the two lib/src/core/threshold/ files, or the
circuits/bin/threshold/user_data_encryption{,_ct0,_ct1}/ packages, which carry READMEs. My
suggestion would be the bin/threshold/user_data_encryption/ package, as the reader-facing one, but
that is a preference rather than something the history decides. Say which you want and I will open it.

circuits/README.md:85 — the wrapper/ section. e651f302 (the new proof-aggregation refactor)
deleted the entire wrapper hierarchy including its 45-line README, and there is no successor document.

This is still not a link fix. The table above the link documents wrapper/dkg/share_computation,
wrapper/dkg/*, wrapper/threshold/*, and a fold circuit, and lines 46-50 explain how
recursive_aggregation/wrapper/dkg/share_computation works — none of which exist under
circuits/bin/recursive_aggregation/ today. Repointing line 85 would turn the link green while
leaving the surrounding section describing an architecture that was replaced. That needs someone who
knows the current aggregation design to rewrite it, so I have left it alone.