v0.2.4

Frontier-autonomy framing: README 'Why this exists for frontier autonomy stacks' (placed up top) + 'for reviewers & safety teams' note. Concept DOI 10.5281/zenodo.20437081.

v0.2.3

RealPage posture correction (code + docs: proposed consent judgment Nov 24 2025, Sherman Act §§1 and 2 — not 'ongoing litigation'); reconciled test/coverage/ADR numbers; six-vertical Autonomy Ladder family update; added .zenodo.json. See CHANGELOG. Archived to Zenodo (concept DOI 10.5281/zenodo.20437081).

v0.2.2 — MI-threshold detector + audit-verify + 51K-execution scrutiny campaign

v0.2.2 — MI-threshold learned-proxy detection + signature-chain verification

Released: 2026-05-28
Citation: Zenodo DOI (minted on this release publish)
Sibling release: linus10x/finserv-agent-audit v1.0.1 (the financial-services pattern library — cre-agent-audit is its CRE-vertical counterpart)

---

What shipped

This release closes two of the three engineering items deferred from the v0.2.1 adversarial-review fold-in. The third — Big-4 named-GC reference quotes (F32) — is research outside the engineering scope and remains open against v0.3.0.

1. Fair-housing MI-threshold learned-proxy detection (closes F11; ADR-0008 update)

A mutual-information detector in fair_housing_preflight.py flags applicant features that carry above-threshold MI against a protected-class reference distribution. Wired through a new FHA-MI-PROXY veto code. Default threshold 0.10 (configurable). Quartile-binning for numeric features; base-2 log entropy normalized by H(Y) so scores read on [0,1].

Distinct from the Module Integrity Proxy that shipped in v0.2.1 under ADR-0013 — that pattern guards the verifier's chain-of-custody. The v0.2.2 detector guards the decision input against learned discrimination through a facially-neutral feature.

Academic anchors: Kusner et al. 2017 (counterfactual fairness), Calmon et al. 2017 (optimized pre-processing), Hardt-Price-Srebro 2016 (equality of opportunity), Pedreshi-Ruggieri-Turini 2008 (discrimination-aware classification).

Settled-liability and ongoing-litigation anchors:

• TransUnion FTC/CFPB consent orders, October 2023, $15M (FCRA § 607(b))
• Louis v. SafeRent Solutions, ~$2.275M class settlement, D. Mass., November 20, 2024 (five-year score-use injunction on voucher-holder applicants)
• U.S. v. RealPage, filed by DOJ and eight state AGs, M.D.N.C., August 23, 2024 (Sherman § 1, ongoing civil antitrust; subsequent data-restrictions milestone November 24, 2025)

A SafeRent-shaped synthetic reference fixture (tests/fixtures/saferent_shaped_reference.py) demonstrates the detector catching zip_code_quintile carrying MI ≈ 0.369 against voucher status with deterministic seed.

2. RFC 3161 signature-chain verification under [audit-verify] extra (closes ADR-0012-A1)

A rfc3161_verify module validates the TSA signature chain on stored RFC 3161 timestamps from audit-ledger entries. Behind the new optional [audit-verify] extra so the package keeps its zero-runtime-dependency posture for callers who only need the issuance path. Install path: pip install "cre-agent-audit[audit-verify]" (pulls cryptography>=42).

3. Pre-DOI deep-scrutiny pass + 51,000-execution property/fuzz campaign

Before this tag, three review chambers (architecture, code-review pass 2, test-strategy + execution proof) ran end-to-end on the post-hardening codebase. The campaign — tests/test_enterprise_scrutiny_campaign.py — runs 18 Hypothesis-powered property and fuzz tests for a total of 51,000 empirically-measured generated examples: MI calculator value-range + permutation invariance, RFC 3161 codec garbage-byte fuzz + request determinism, AuditEntry canonical-bytes determinism + frozen contract, AuditLedger verify-chain invariant + tamper detection + chain-head consistency, ProtectedClassReference paired rejection + acceptance, MIThresholdDetector finding invariants, quartile-bin + severity-ladder monotonicity, and full round-trip across the three LedgerStore backends. The deep-review pass also added @runtime_checkable parity to the TimestampSource and WitnessRegister Protocol seams, threat-model framing to the RFC 3161 codec docstring, and Python memory-posture documentation to LocalMIProxy.from_env.

Engineering posture preserved

• Zero runtime dependencies on the default install path
• cryptography>=42 only when [audit-verify] is requested
• 336 tests pass · 86% branch coverage · ruff + mypy --strict clean
• 9 core patterns + 4 hardening patterns + 1 category-claim pattern (ADR-0014: operator-side AI governance)

Verification

pip install -e ".[dev,audit-verify]"
pytest -q
ruff check src tests
mypy --strict src tests

Compatibility

v0.2.2 is a non-breaking addition to v0.2.1. The detector is opt-in (default None on FairHousingPreflightGate); the extra is opt-in. Callers who do not import either path see no behavior change.

Roadmap

v0.3.0 (target 2026-Q4): full per-pattern ISO/IEC 42001 mapping at subordinate-clause level; five state regulatory mappings; pruning unused agents/ stubs; LangChain + CrewAI adapters; Docker compose for a 60-second zero-pip-install demo. See ROADMAP.md.

Provenance and disclaimers

cre-agent-audit is a reference architecture, not legal advice. Every regulatory citation in this release is primary-source-verified. See DISCLAIMER.md and LIMITATIONS.md for posture and caveats.

Sibling repository: linus10x/finserv-agent-audit.

v0.2.1 — Zenodo Archive

This release enables permanent archiving of the v0.2.0 — Foundation contents to Zenodo. Same source. Same nine MIT-licensed governance patterns for AI in commercial real estate operations. Same anchors to three named CRE-AI regulatory matters of record (In re Trans Union Rental Screening Solutions joint FTC/CFPB consent orders October 2023, Louis v. SafeRent Solutions LLC class settlement November 2024, U.S. v. RealPage Inc. filed by DOJ + 8 state AGs August 23 2024). Same zero-runtime-dependency, mypy --strict clean codebase. Same 142 tests + 89% branch coverage.

What's new in v0.2.1

• Zenodo GitHub integration enabled for this repository
• This release is the first archived to Zenodo with a permanent DOI
• A Zenodo DOI badge will land in the README in a follow-up commit

The intent: when a board member, regulator, or peer reviewer asks "where is the primary source?" — the answer is a Zenodo-minted DOI, not a mutable GitHub README.

Companion repository linus10x/finserv-agent-audit gets the same archiving treatment in v1.0.1.

---

License: MIT — fork freely; no warranty.
Citation: see CITATION.cff. Zenodo DOI minted on this v0.2.1 release.
Trademark: Autonomy Ladder™ is a common-law trademark of Kunjar Bhaduri; USPTO registration planned in classes 9, 35, 41, 42.

v0.2.0 — Foundation

v0.2.0 — Foundation

First public release of cre-agent-audit. Nine MIT-licensed governance patterns for AI in commercial real estate operations, built to a single design philosophy: durable artifacts, not slideware.

What this is

A reference architecture for AI-governance discipline inside CRE operating companies. Nine patterns covering DEFCON state, Sovereign Veto, hash-chained Audit Ledger, Autonomy Ladder A0→A4, Regulation Mapping, Shadow-Mode Rollout, Lease-Abstraction Provenance, Fair-Housing Pre-Flight Gate, and Tenant PII Data Residency. Two additional design ADRs (ADR-0010 audit-chain retention/privilege/discovery posture; ADR-0011 vendor-output adapter pattern) cover the layered policy and vendor-mediated AI surface.

Every pattern produces an artifact — code, ADR, audit-ledger entry, veto-config, control-description table, vendor-clause template — that survives outside the engagement that created it. Patterns are MIT-licensed, zero-runtime-dependency (stdlib only; YAML is author-time only via scripts/build_compliance_json.py), and mapped to primary-source regulatory citations.

The repo is meant to compound: fork it, ship it inside your stack, copy patterns into your own audit framework, cite the ADRs in your risk register. The patterns are non-proprietary; the implementation is yours.

Settled-liability anchors (primary-source verified)

The patterns are designed against three regulatory matters in the CRE-AI surface:

• In re Trans Union Rental Screening Solutions — joint FTC/CFPB consent orders, October 2023, $15M — FCRA § 607(b) accuracy in rental-screening reports
• Louis v. SafeRent Solutions, LLC, No. 1:22-cv-10800 (D. Mass.) — November 2024 — approximately $2.275M class settlement; settlement included a five-year score-use injunction on voucher-holder applicants
• U.S. v. RealPage, Inc. — filed August 23, 2024 by DOJ + 8 state AGs — ongoing civil antitrust litigation alleging algorithmic rent-coordination

Doctrinal foundation: Texas Dept. of Housing v. Inclusive Communities Project, 576 U.S. 519 (2015).

Install

git clone https://github.com/linus10x/cre-agent-audit.git
cd cre-agent-audit
pip install -e ".[dev]"
make verify                                            # full gate
python examples/02_tenant_screening_preflight/run.py   # demonstrates FHA-PROXY/VOUCHER/SOI/CRIM/DISPARATE

Cold-clone to verified output target: under 60 seconds (warm pip cache). See docs/REPRODUCE.md.

What's in v0.2.0

• 9 governance patterns + 142 unit tests + 89% branch coverage + ruff + mypy --strict clean
• Zero runtime dependencies (preserves parity with the sibling linus10x/finserv-agent-audit)
• 11 ADRs (9 pattern primitives + 2 layered policy ADRs)
• 9 per-pattern Control Description Tables (docs/controls/)
• Four-framework mapping matrix: NIST AI RMF × ISO/IEC 42001 × COSO ICAIR × Big-4 standard taxonomy (docs/MAPPING-MATRICES.md)
• 3 drop-in vendor-clause templates (docs/vendor-clauses/{screening,abstraction,pricing}.md)
• PE operating-partner due-diligence 10-question checklist (docs/PE_DUE_DILIGENCE.md)
• 90-day deployment cadence walkthrough framed as privileged engineering rails (examples/FIRST_90_DAYS.md)
• Three FINOS-format contributory control drafts with explicit non-endorsement provenance (governance-artifacts/)
• DISCLAIMER.md + LIMITATIONS.md + PRIOR-ART.md
• Sibling-parity files: CITATION.cff, CODE_OF_CONDUCT.md, ROADMAP.md, .pre-commit-config.yaml, .github/CODEOWNERS, FUNDING.yml, dependabot.yml, ISSUE_TEMPLATE, PR template

Sibling

Maps 1:1 with linus10x/finserv-agent-audit for financial-services workflows. One framework, two named verticals, one author.

Acknowledgements

NIST AI Risk Management Framework · Treasury Financial Services AI Risk Management Framework · FINOS AI Risk Initiative · Marcos López de Prado (named advisor on adjacent work) · Solon Barocas + Moritz Hardt + Arvind Narayanan (Fairness and Machine Learning foundational text) · Andrew Selbst + Danah Boyd + Sorelle Friedler + Suresh Venkatasubramanian + Janet Vertesi (Fairness and Abstraction in Sociotechnical Systems, FAT* 2019) · Margaret Mitchell et al. (Model Cards for Model Reporting) · Timnit Gebru et al. (Datasheets for Datasets) · Inioluwa Deborah Raji et al. (Closing the AI Accountability Gap).

Notice

This repository is a reference architecture, not legal, regulatory, audit, or fairness-testing advice. Regulatory characterizations are summaries; readers must consult qualified counsel for jurisdiction-specific compliance. No attorney-client relationship is formed by use of this repository. See DISCLAIMER.md.

What's next (v0.2.1 — adversarial-review follow-ups; target 2026-Q3)

• Implement MI-threshold learned-proxy detection in the Fair-Housing Pre-Flight Gate (v0.2.0 ships lexical-only with a bounded ADR-0008 claim)
• Pluggable persistence backend for the Audit Ledger
• RFC 3161 trusted-timestamp integration
• OpenTimestamps / Sigstore Rekor witness-anchor reference integration
• VendorScoreGate concrete implementation (v0.2.0 ships ADR-0011 design)
• Full negative-results / failure-mode appendix
• Named-GC reference quotes

See ROADMAP.md for the full v0.2.1 + v0.3 + v0.4 horizon.