[codex] address release governance and audit findings

[Fieldnote-Echo]

Summary

• document and audit the two-approver crates-io / pypi GitHub Environment gate: Fieldnote-Echo + toadkicker, self-review prevention, 30-minute wait timer, and stable release tag policy
• expose the calibration-profile byte limit in the ordvec-manifest Python bindings, default resource limits, exports, and verifier/create keyword arguments
• align .ovfs / OVFS security and provenance docs with the public FastScan loader and fuzz target
• refresh threat-model maintenance-budget language for the two-approver model and clarify the C ABI empty-candidate contract
• hide the SubsetScratch::capacities_for_test helper behind test-utils, fix stale FastScan/docs.rs and b=8 rustdoc wording, and remove a duplicate Python doc paragraph
• update formalization links and release invariants after ordvec-formalization moved under Project-Navi

Why

A pre-release audit found one published Python manifest API parity gap, several stale security/governance/provenance docs, and small API/docs hygiene issues. The live release environments and transferred formalization repo also needed the checked-in release surfaces to match current external state.

Validation

• bash -n tests/release_environment_settings.sh
• bash tests/release_environment_settings.sh
• cargo fmt --check
• cargo check -p ordvec-manifest-python --locked
• cargo test -p ordvec-manifest-python --locked
• cargo clippy -p ordvec-manifest-python --all-targets --locked -- -D warnings
• maturin build --release --manifest-path ordvec-manifest-python/Cargo.toml --out /tmp/ordvec-manifest-python-dist
• uv run --with pytest --with /tmp/ordvec-manifest-python-dist/ordvec_manifest-0.5.0-cp310-abi3-manylinux_2_34_x86_64.whl python -m pytest ordvec-manifest-python/tests -q
• cargo test --locked
• cargo test --locked --features test-utils batched_into_is_allocation_free_after_warmup
• python3 tests/release_publish_invariants.py
• cargo clippy --all-targets --all-features --locked -- -D warnings
• cargo clippy -p ordvec-python --all-targets --locked -- -D warnings
• cargo build -p ordvec-ffi --locked
• cargo test -p ordvec-ffi --locked
• git diff --check

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

[qodo-code-review]

PR Summary by Qodo

Document dual-approver release environment gates and audit checks
📝 Documentation ⚙️ Configuration changes 🧪 Tests 🕐 40+ Minutes

Description

• Document two-approver crates-io/pypi environment protections (no self-review, 30-minute wait).
• Update the pre-tag environment audit to assert reviewer set, wait timer, and tag policy.
• Refresh threat model residual risk from single-approver compromise to compromise/collusion.

Diagram

graph TD
  A["Tag push (vX.Y.Z)"] --> B["GitHub Actions: release.yml"] --> C["Guard checks"] --> D{"Env gate: crates-io"} --> F["Publish to registries"]
  C --> E{"Env gate: pypi"} --> F
  G["Environment audit script"] --> D
  G --> E

Loading

High-Level Assessment

The following are alternative approaches to this PR:

1. Use a GitHub Team as required reviewer (vs individual accounts)

• ➕ Easier rotation/onboarding without doc/script changes
• ➕ Avoids coupling release policy to specific usernames
• ➖ Requires extra team governance and careful membership management
• ➖ May reduce individual accountability if team is too broad

2. Manage environment protections as IaC (GitHub API/Terraform)

• ➕ Makes the critical environment settings drift-resistant and reviewable in PRs
• ➕ Can enforce reviewer set, no-self-review, and wait timer automatically
• ➖ Adds tooling/credentials complexity
• ➖ Still limited by GitHub’s partial support for environment-as-code features

3. Split publishing into a separate manually-dispatched workflow

• ➕ Clearer separation between build/attest and publish authorization
• ➕ Manual dispatch can enforce different permission boundaries
• ➖ More operational steps and potential for process errors
• ➖ Doesn’t inherently replace environment-level protections for OIDC minting

Recommendation: The PR’s approach (documenting the environment protections and strengthening the audit script) is the right baseline because GitHub Environment settings remain out-of-band from workflow code. Consider switching required reviewers to a tightly-scoped GitHub Team if reviewer rotation becomes common; otherwise the current explicit dual-approver documentation is clear and auditable.

Files changed (7) +95 / -54

Tests (1) +23 / -9

release_environment_settings.shAudit environments for dual reviewers, no self-review, and wait timer +23/-9

Audit environments for dual reviewers, no self-review, and wait timer

• Extends the GitHub API audit to assert two required reviewers (sorted), 'prevent_self_review=true', presence/value of a wait_timer rule (30 minutes), and keeps enforcing the stable tag pattern policy.

tests/release_environment_settings.sh

Documentation (5) +66 / -40

CHANGELOG.mdAdd security note about dual-approver environment gates +9/-0

Add security note about dual-approver environment gates

• Introduces an '[Unreleased]' Security entry documenting the two required reviewers, self-review prevention, and 30-minute wait timer for registry publishing environments.

CHANGELOG.md

CONTRIBUTING.mdDocument release gating behavior for contributors +4/-3

Document release gating behavior for contributors

• Updates contributor-facing release notes to specify the two required reviewers, self-review prevention, and the 30-minute environment wait timer before registry publishes proceed.

CONTRIBUTING.md

GOVERNANCE.mdRecord environment protections as a governance control +4/-0

Record environment protections as a governance control

• Adds governance-level documentation that release publishes are gated by 'crates-io'/'pypi' environments with two approvers, self-review blocked, and a wait timer.

GOVERNANCE.md

RELEASING.mdExpand release procedure to match current environment protections +29/-17

Expand release procedure to match current environment protections

• Rewrites the environment protection section to include dual reviewers, prevent self-review behavior, and the 30-minute wait timer; updates the pre-tag audit checklist and operator guidance for approvals.

RELEASING.md

THREAT_MODEL.mdUpdate supply-chain residual risk for dual-approver release gates +20/-20

Update supply-chain residual risk for dual-approver release gates

• Reframes THREAT-SUPPLY-001 residual from single-maintainer compromise to approver compromise/collusion and updates the threat table and roadmap notes accordingly.

THREAT_MODEL.md

Other (1) +6 / -5

release.ymlClarify manual environment gate requirements for publish jobs +6/-5

Clarify manual environment gate requirements for publish jobs

• Updates release workflow comments and environment annotations to reflect the dual-approver gate, prevent-self-review setting, and 30-minute wait timer for 'crates-io' and 'pypi' environments.

.github/workflows/release.yml

[qodo-code-review]

Code Review by Qodo

🐞 Bugs (0) 📘 Rule violations (0) 📎 Requirement gaps (0)

Great, no issues found!

Qodo reviewed your code and found no material issues that require review