Scaffold circuit_unpacker crate

[anatgstarkware]

Summary

New circuit_unpacker crate at crates/circuit_unpacker/, added to the workspace members. This is the circuits-DSL analogue of the Cairo unpacker bootloader (BLAKE_UNPACKING mode): it walks a statically-shaped perfect-binary-tree recursion of circuit_verifier proofs and emits each real leaf's outputs.

This PR is scaffolding only — documented public API with a todo!() body:

• UnpackerHints — prover-supplied per-real-leaf outputs; derives tree shape from leaf_outputs.len().
• DUMMY_LEAF_MARKER (0xDEAD) and dummy_leaf_output() — canonical padding for non-power-of-two N.
• run_unpacker() — entry point, todo!() for now.

Tree shape is a complete perfect binary tree fixed by N; right-aligned slots are dummies. No n_subtasks / size headers in preimages — positions are compile-time-known.

---

This change is 

[cursor]

PR Summary

Low Risk
No production wiring or cryptographic enforcement yet—only API, docs, and a placeholder implementation plus a minimal validity test.

Overview
Adds a new workspace crate circuit-unpacker (crates/circuit_unpacker/) as the circuits-DSL counterpart to the Cairo bootloader’s Blake unpack path: it will walk a fixed perfect-binary-tree of recursive circuit_verifier proofs and surface each real leaf’s outputs via Blake bindings aligned with circuit_verifier::verify.

This change is scaffolding only: extensive module docs, UnpackerHints (tree shape derived from leaf_outputs.len()), dummy padding via DUMMY_LEAF_MARKER / dummy_leaf_output(), and run_unpacker which currently returns empty per-leaf Vecs with a TODO instead of building constraints. A starter test unpacker_n2_no_dummies builds an off-circuit root hash and only asserts ctx.is_circuit_valid() until the real tree walk lands. Cargo.toml / Cargo.lock register the new member.

^{Reviewed by Cursor Bugbot for commit 507d8f1. Bugbot is set up for automated code reviews on this repo. Configure here.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 59.39%. Comparing base (c0b937b) to head (507d8f1).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #356      +/-   ##
==========================================
+ Coverage   59.19%   59.39%   +0.20%     
==========================================
  Files          35       36       +1     
  Lines        5303     5330      +27     
==========================================
+ Hits         3139     3166      +27     
  Misses       2164     2164              

Files with missing lines	Coverage Δ
crates/circuit_unpacker/src/lib.rs	100.00% <100.00%> (ø)

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, have a team admin enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 951f194. Configure here.}

[cursor]

Stub silently returns empty Vec instead of panicking

Medium Severity

run_unpacker returns Vec::new() instead of calling todo!(). The PR description explicitly states this function has a "todo!() body," but the actual implementation silently returns an empty vector. Any caller (including future integration tests or downstream crates) will receive an empty result without any runtime indication that the function is unimplemented, potentially masking real bugs. The documented soundness contract in the docstring is quietly violated rather than loudly failing.

 

^{Reviewed by Cursor Bugbot for commit 951f194. Configure here.}