# Changelog

> **This is the PUBLIC form of this artifact. Findings are listed in full, but the details of SECURITY findings — which rule fired, in which file, on which line, and how to fix it — are deliberately withheld, and any secret-scanner results are excluded entirely. Where detail is absent here it was REMOVED FOR PUBLICATION; it is not missing from the analysis. The complete artifact is available from the repository owner.**

## Score

- CAI 47 → 46 (-0.3)
- Rubric changed (rubric-2026.10.5 → rubric-2026.11.2) — scores are not directly comparable.

## Lenses

- Code Health 42 → 42 (+0.0)
- Architecture 91 → 73 (-17.5)
- Maturity 63 → 63 (-0.2)
- Readiness 40 → 40 (+0.4)
- Security 60 → 63 (+2.7)
- Domain Modelling 100 → 100 (+0.0)
- Accessibility 50 → 48 (-2.1)
- Performance 91 → 91 (+0.0)

## Resolved (9)

- Ambiguous internal vs public API methods. `apply` and `apply_r` exist in multiple places. `apply_r` appears in `Core`, `CachedApply`, and `Instrumentation`. The suffix `_r` is not explained and likely stands for 'raw' or 'recursive', but it is exposed in the public surface without clear distinction from `apply`. This suggests an inconsistency in how the core execution logic is exposed.
- Duplicate intent with different signatures/contexts. `Behaviour.allowed_to?` is a class/module-level check, while `Policy.Core.allowed_to?` is an instance-level check. While technically distinct, the naming collision with `allowed_to?` on the policy instance (which usually implies checking the current policy's state) vs the behaviour (checking against a record) is confusing. Furthermore, `Policy.Core` has `apply(rule)` which is the primary execution method, making `allowed_to?` on the instance somewhat redundant or ambiguous compared to `apply`.
- High CVE: [GHSA redacted] (tutorial/package-lock.json)
- High CVE: [GHSA redacted] (tutorial/package-lock.json)
- High CVE: [GHSA redacted] (tutorial/package-lock.json)
- Medium CVE: [GHSA redacted] (tutorial/package-lock.json)
- Medium CVE: [GHSA redacted] (tutorial/src/templates/default/package-lock.json)
- Off-boarding risk: anonymized user #1
- ResponseOutparam.finish (cognitive 20) (tutorial/src/templates/default/lib/server.js)

## New (7)

- High CVE: [CVE redacted] (tutorial/package-lock.json)
- High CVE: [GHSA redacted] (tutorial/src/templates/default/package-lock.json)
- High CVE: [GHSA redacted] (tutorial/package-lock.json)
- High CVE: [GHSA redacted] (tutorial/package-lock.json)
- High: security finding (details withheld)
- High: security finding (details withheld)
- Off-boarding risk: anonymized user #1

## Changes since last survey

- 12 commits — 10 feature/other, 2 fixes

## By area

- (root) — 4 commits
- test/action_policy — 3 commits
- lib/action_policy — 2 commits
- (repo) — 1 commit
- benchmarks/authorization_hot_path.rb — 1 commit
- tutorial/ruby-wasm — 1 commit

## Notable commits

- fix: - fix tutorial's gemfile.lock
- fix: Fix policy lookups for partially selected Active Record models
- change: * cache: calculate cache_key once per policy resolution
- change: * forspell -> rubocop-spellcheck
- change: + SECURITY.md
- change: - cache: avoid collisions in thread memoized store
- change: - ext: remove import_methods
- change: - tests: rails 8.2 compat
- change: Add a benchmark for authorizing Active Record models from a controller
- change: Bump 0.7.8
- change: Keep cache_key_with_version, key only partially selected records by identity
- change: Merge pull request #314 from czj/perf/active-record-policy-cache-key

## Architecture

- Unchanged — 0 containers · 1 contexts · 0 edges
