Institutional governance for commercial real estate
127 CRE skills. One control plane. Four-eyes approval on everything that reaches a lender.
CRE Skills Pro is the governance layer for institutional teams using Claude Code. Every output that leaves the deal room, from an IC memo to a lender package to an LP report, passes a configurable four-eyes approval matrix before it can be written or exported. The audit trail comes with it. It works with the free CRE Skills plugin your team already has.
Every stage of the deal has a gate
Built for the whole fund, not just the deal PRO PREMIUM
Twenty fund and portfolio-level skills for institutional managers, on top of the 127 CRE skills. This is the surface a large asset manager buys on.
Fund risk and treasury
- portfolio-risk-aggregator
- fund-liquidity-forecaster
- interest-rate-hedge-manager
- nav-facility-analyzer
Valuation governance
- valuation-committee-package
- appraisal-reconciler
- fair-value-bridge
- portfolio-stress-test
LP relations and IR
- capital-account-statements
- lp-ddq-response-engine
- side-letter-mfn-tracker
- co-investment-allocation
Regulatory and compliance
- form-pf-adv-assistant
- marketing-rule-checker
- conflicts-register
- books-records-audit-kit
Tax, structure, and ESG
- reit-compliance-tester
- ubti-eci-blocker
- esg-gresb-builder
- ilpa-reporting-builder
Asset management and diligence
- lease-abstraction-pro
- loan-abstraction-pro
- property-condition-assessment-pro
- offering-memorandum-pro
- closing-documents-pro
What the control plane gives you
Four-eyes matrix
An author cannot approve their own IC memo or LP capital statement. The matrix is keyed to deal size, output type, and role.
Append-only audit
Every generate, approve, reject, export, and override, with the actor, the timestamp, and the artifact hash.
Setup wizard
Eleven screens that configure your org profile and thresholds, then run a simulated $84M deal so you can see which gates fire and who is summoned.
Verified output, not just generated output PRO LOOP HARNESS
The Pro Loop Harness turns a skill's output into a verified artifact. The load-bearing CRE numbers are recomputed from the source rows by a checker that is separate from the model that produced them, and every artifact is labeled by how far it has been verified.
Deterministic verifiers
NOI yield, DSCR, WALT, occupancy, and the debt stack are recomputed from the source rows, not trusted from the model. A simple average dressed up as a weighted rollup, a dropped row, a stale date, or a unit drift cannot pass.
The maker never grades itself
A separate verifier produces the verdict, so the number you ship was checked by something other than the thing that wrote it. A non-finite or unreconciled value fails closed.
A maturity label on every artifact
Each output is labeled by how far it has been verified. Nothing self-labels decision-grade; that takes reconciliation, provenance, a verifier pass, and a named human approval.
| Maturity label | What it requires |
|---|---|
| draft / advisory | may be produced by a loop |
| verified_local | deterministic checks pass, all verifiers green |
| human_approved | a named approver in the four-eyes ledger |
| decision_grade | source grade, reconciliation, provenance, a verifier pass, and a named human approval |
We tell you what governance means here
In self-install mode you get strong session-level enforcement and a tamper-evident audit log. Operator-proof enforcement, where an approval token cannot be self-satisfied, needs managed deployment plus a server-side identity and approval service. Your contract states which mode you run. Both are real governance. Read the guarded vs enforced disclosure.
Pricing
Free
- 127 CRE skills
- Routing and calculators
- No governance or audit
Pro Individual
- Local governance hooks
- Personal local audit log
- Setup wizard and private catalog
Pro Team
- Four-eyes approval matrix and shared audit
- 20 institutional skills (Pro only)
- Multi-user deal-state memory
- Managed-settings push
Enterprise
- Server-side enforced four-eyes
- Immutable WORM audit
- SSO and SCIM, SOC 2, VPC option
See full pricing and the comparison table. Prices are early hypotheses.
Procurement-ready: SOC 2 in progress, managed-settings deployment, a vendor questionnaire, an air-gapped option, and an MSA and DPA on request. Download the security package.