Plugin repository for the CRE-150 training workshop in the 021 Events series.

CRE-150 covers applied AI workflows for commercial real estate professionals using Claude Cowork. These plugins are the hands-on toolkit for the session — participants install them into their Cowork workspace and run the live extraction exercises against real lease documents and MLS reports.

Abstracts commercial real estate leases using the REIXS-LA-NA-001 extraction standard and a 258-field Domain Data Dictionary across 25 sections.

Trigger: "Abstract this lease" + point to a PDF or DOCX

Default mode — Full Abstraction

Extracts all 25 DDD sections with FACT / INFERENCE / MISSING / CONFLICT status tagging, AutoFail guards, Schedule G override detection, and financial risk analysis. Output: Markdown (30–40 KB) or JSON (50–60 KB), saved to your workspace Reports/ folder.

--criticaldates mode — Critical Dates Calendar

Extracts every date-triggered obligation across 10 categories, assigns P1–P4 priority, and calculates cascading reminders. Optional --ics and --csv output for import into Outlook, Google Calendar, or lease management software.

Abstract this lease /path/to/lease.pdf
Abstract this lease /path/to/lease.pdf --criticaldates --ics --csv
Abstract this lease /path/to/lease.pdf -json

Extracts commercial MLS property listings from PDF reports into professionally formatted Excel spreadsheets.

Trigger: "Extract MLS data from" + point to a PDF

Reads the PDF with LLM vision, extracts a 34-field schema for every listing, verifies the extracted JSON against the source PDF, then produces an Excel file with subject property highlighting and correct column ordering.

Extract MLS data from Mississauga_industrial.pdf
Extract MLS data from report.pdf --subject="2550 Stanfield"

MCDA ordinal ranking for fee simple sales comparison valuation. Ranks a subject property and comparables on weighted characteristics, maps composite scores to value via interpolation and regression.

Trigger: "Run MCDA sales comparison" + point to a PDF or JSON

Accepts CoStar or broker comparable sale reports, extracts property attributes, calculates a composite MCDA score for each sale, and interpolates a value opinion for the subject property. Supports up to 25 weighted variables with dynamic weight redistribution when data is sparse.

Run MCDA sales comparison /path/to/comparables.pdf
Run MCDA sales comparison /path/to/input.json --stats

MCDA competitive positioning analysis for commercial real estate leasing. Ranks a subject property against market comparables on up to 25 weighted variables and provides strategic pricing recommendations to achieve Top 3 market positioning.

Trigger: "Relative valuation" / "competitive positioning" / "rank this property" + point to a PDF or JSON

Extracts property data from CoStar or broker market reports, calculates driving distances (optional, requires Distancematrix.ai API key), runs the MCDA ranking engine, and produces a sensitivity analysis showing exactly what it would take to reach the Top 3 competitive tier.

Supports four tenant personas — Default, 3PL/Distribution, Manufacturing, Office/Flex — each with pre-tuned variable weights.

/mcda-lease-comparison /path/to/market_report.pdf
/mcda-lease-comparison /path/to/input.json --persona 3pl --full --stats

Tenant credit analysis for commercial real estate lease approvals. Extracts financial data from PDF financial statements, runs a 100-point weighted credit scoring algorithm (A–F rating), estimates default probability and expected loss, and generates a comprehensive credit report with security deposit and approval recommendations.

Trigger: "Tenant credit analysis" / "analyze tenant financials" / "creditworthiness" + point to a PDF

Accepts one to three years of financial statements. Extracts balance sheet and income statement data, calculates 15+ financial ratios across four categories (liquidity, leverage, profitability, rent coverage), scores the tenant on a 100-point scale, and recommends a security amount with step-down schedule.

/tenant-credit /path/to/2024_financials.pdf
/tenant-credit /path/to/2024_financials.pdf /path/to/2023_financials.pdf
/tenant-credit /path/to/2024_financials.pdf /path/to/lease_proposal.pdf

Landlord-side CAM true-up for commercial real estate. Combines lease-aware recoverability rules with deterministic Python math so the same inputs always produce the same tenant charges, landlord absorption totals, and comparison output.

Trigger: /cam-reconcile <property-dir>

Ingests a property package (property.yaml, leases.json, gl.csv, budget.md), applies lease-aware corrections for duplicate invoices, turnover charges, management-fee basis errors, modified-gross exclusions, restaurant carve-outs, base years, and CAM caps, then emits typed manifests, tenant PDF statements, a workpaper workbook, an audit log, and an optional Anthropic comparison report.

/cam-reconcile <property-dir>
/cam-explain <allocated-manifest.json> <tenant-id-or-line-id>
/cam-compare-vs-anthropic <allocated-manifest.json> <anthropic-output.txt>

Owner-side tender evaluation for Ontario roof replacement projects. Scores contractor bids against the owner's RFP using mandatory pass/fail gates plus a weighted MCDA rubric. Supports commercial (OBC Part 3) low-slope membrane systems and residential (OBC Part 9) steep-slope systems.

Trigger: /roof-review <rfp.pdf> <bid1.pdf> <bid2.pdf> ...

Parses the RFP and each bid into a normalized tender manifest, applies qualification gates (WSIB, CGL, bonding, WAH training, Skilled Trades Ontario, financial capacity), scores a weighted matrix across Price / Technical / Experience / Warranty / Schedule / Qualifications, and produces a scoring matrix, a red-flag report, and a recommendation memo with citations to OBC, CRCA, and CCDC-23.

/roof-review <rfp.pdf> <bid1.pdf> <bid2.pdf> ...
/roof-redflags <manifest.json>
/roof-memo <manifest.json>

Lease deal investment analysis using the Ponzi Rental Rate (PRR) framework (Chan, 2015). Extracts terms from lease documents, matches landlord investment parameters from a configurable database, calculates NER/GER/NPV, determines breakeven thresholds at four levels, and generates an investment recommendation report.

Trigger: "Analyze this lease deal" / "Calculate effective rent" / "NER analysis" + path to lease document + path to landlord parameters JSON

Reads the lease or offer to lease (PDF, DOCX, MD), matches the landlord named in the PARTIES section against landlord_investment_parameters.json, builds a validated input JSON, runs eff_rent_calculator.py, and produces a Markdown report with NPV breakdown, NER vs breakeven comparison, and APPROVE / NEGOTIATE / REJECT recommendation.

/effective-rent /path/to/offer.pdf /path/to/landlord_params.json
/effective-rent /path/to/lease.pdf /path/to/landlord_params.json /path/to/ti-quote.pdf

All eight plugins are designed specifically for Claude Cowork's context architecture. Each skill follows a three-step dispatch pattern:

1. Primary context (thin) — resolves file paths and environment variables, dispatches a subagent
2. Extraction subagent (full 200k window) — handles all document reads, extraction, and file writes
3. Primary context — relays the structured result to the user

This keeps the primary context from being exhausted by large document reads, which is the most common failure mode for multi-step skills in Cowork. See cowork_lessons_learned.md for the full design rationale.

Load all plugins into your Claude Cowork workspace by pointing to this repository in your plugin settings.

• cre-lease-abstraction, mls-extractor, mcda-sales-comparison, mcda-lease-comparison, tenant-credit, cam-reconciliation-cre, roof-replacement-review — no API keys required
• mcda-lease-comparison — optional: set DISTANCEMATRIX_API_KEY for automatic driving distance calculation (free tier: 1,000 elements/month at distancematrix.ai)
• cam-reconciliation-cre, roof-replacement-review — require Python dependencies from each plugin's requirements.txt

Series: 021 Events Session: CRE-150 — Claude Cowork for Real Estate Author: Reggie Chan