More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    TerraVue — US Property Analysis

    1

    US housing market simulator using Monte Carlo paths and FHFA data to solve the buy vs. rent dilemma.

    Free

    1 useSecurity scanned
    TerraVue — US Property Analysis

    TerraVue — US Property Analysis

    Example session with this skill installed

    Should I buy a $425,000 house in Minneapolis 55408 if my rent is $1,900 and I plan to stay 7 years? Household income is $140,000.

    • Read your context and instructions
    • Compiled the terravue

    **Short answer:** you can afford it, but on these numbers renting wins over 7 years — by a wide margin. | | Result | | :--- | :--- | | Affordability | **Comfortable** — 24.6% DTI, clears the $122.9K income bar by $17K | | Cash to close | $95,625 (20% down + ~$10.6K closing) | | Year-1 monthly cost | **$3,222 buying** vs. $1,920 renting (+$1,300/mo) | | Breakeven | None within 7 years | **Buy vs. rent at year 7 (500 Monte Carlo scenarios, 6.75% rate)** | Renter invests at… | P(buying wins) | Median net-worth gap | Best-case (p90) | | :--- | :--- | :--- | :--- | | 9.5% (equities) | **3%** | −$167K | −$55K | | 4.5% (savings/bonds) | **15%** | −$109K | +$18K | ### Why it loses - **Weak appreciation forecast:** 55408 is modeled at ~1.3%/yr — barely enough to cover selling costs on exit. - **Price-to-rent ~18.6×:** $1,900 rent is cheap relative to a $425K purchase. - **Carry costs:** 1.02% property tax, 0.85% insurance, plus maintenance — ~$1,000/mo of non-equity cost on top of mortgage interest. ### What would flip it - Your $1,900 is rent for a *comparable* home. If an equivalent 55408 rental is really ~$2,600+, the math tightens a lot — worth checking. - A materially lower price, a rate in the low-5s, or a much longer hold (10+ yrs). - Strong non-financial reasons (stability, control of the space) you're willing to pay ~$100K+ in expected net worth for. Want me to run a sensitivity on the rent comp or a 10-year hold?

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Buying a home is often the largest financial decision a person makes, yet most advice relies on vague estimates or calculators that apply a national average to a local decision. This skill replaces guesswork with a 500-scenario Monte Carlo simulation grounded in FHFA house-price records at ZIP granularity back to 1975, plus Zillow ZORI rents and Census ACS data. ## What it does - **Buy vs. Rent Analysis** returns the probability that buying outperforms renting over a specific horizon, modeling the renter investing the down payment and any monthly difference so both sides are compared on the same money over the same clock. - **Affordability Modeling** determines whether a specific property fits a buyer's profile, or solves for the maximum purchase price under lender DTI limits and cash on hand — and reports which limit actually binds. - **Location Benchmarking** retrieves ZIP-level appreciation estimates, school ratings, tax and insurance rates, walkability, water quality and market rents. - **Comparative Analysis** ranks multiple ZIPs side by side on appreciation, schools, education level and water quality. - **Breakeven Estimation** reports the year buying overtakes renting — or states plainly that it does not within the horizon, which is frequently the honest answer. ## How it works 1. **Gather context** by identifying intent — evaluating a specific address, or determining a general budget. 2. **Retrieve local data** from the address or ZIP to pull regional tax, insurance, rent and appreciation history. 3. **Run simulations** across 500 scenarios, sampling from each ZIP's own historical volatility rather than a national figure. 4. **Report probabilities** rather than verdicts, along with the assumptions driving them. ## Frameworks & tools A remote MCP server at https://api.terravue.app/mcp — no install, no runtime, no API key. It draws on FHFA HPI, Zillow ZORI and Census ACS. ## Why this beats prompting it yourself Language models reliably hallucinate mortgage math and have no access to ZIP-specific appreciation history. This skill forces the agent to call a dedicated engine and verified datasets instead of producing plausible-looking but incorrect figures — and, just as importantly, to refuse the questions the data cannot answer. ## Use cases - Determine whether an $800k home in Austin beats renting over a 5-year hold. - Identify whether a buyer is limited by monthly income (DTI) or by liquid cash. - Rank several ZIP codes against each other on appreciation and school ratings. - See how sensitive a decision is to the assumed investment return, by running it at 9.5% and again at 4.5%. ## Known limitations US residential properties only. No individual property valuations, no listing lookups, and no sub-ZIP neighborhood forecasting — sub-ZIP appreciation edges mean-revert, so census-tract data describes where a home sits within its ZIP today and never predicts that pocket's future.

    US residential properties only. No international coverage, no commercial property. Figures are ZIP-level. The tool describes a market, not a specific property: it will not produce a valuation for an individual house, and it will not fetch or look up listings. If you know your actual rent or an actual asking price, pass it in — it's an input, not a lookup. No neighborhood-level forecasting. Sub-ZIP appreciation edges mean-revert: past outperformance correlates about -0.3 with future outperformance. Census-tract data is therefore reported as where a home sits within its ZIP today, and never as a prediction about that pocket. The appreciation figure is an estimate under a stated model, not a prediction. The Monte Carlo returns a probability under assumptions you can change — most significantly a 9.5% default return on money not spent on the home, which is the input that moves results the most. Not financial advice. It is a simulator with stated, overridable assumptions.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    1 use

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 29 days ago

    • Free with an Agensi account

    Permissions

    Network Access

    Network access only, to reach the remote server at api.terravue.app. The package is documentation — SKILL.md, a config template and install instructions. It runs no code locally, reads and writes no files, and needs no credentials or environment variables.

    Listed29 days ago

    What's inside

    Frequently Asked Questions

    A standard calculator applies national or metro-level averages and returns a single confident number. This runs FHFA repeat-sales house price records at ZIP granularity back to 1975 through 500 Monte Carlo scenarios, so you get a probability and a breakeven year when one exists instead of a point estimate. It also models the renter investing the down payment and any monthly difference, so both sides are compared on the same money over the same clock.

    Nothing. TerraVue is a remote MCP server at https://api.terravue.app/mcp — no install, no runtime, no API key, no account. Any client supporting streamable-http MCP works, including Claude Code, Claude Desktop, Cursor and VS Code Copilot. Claude Code is one command. Clients that only speak stdio can bridge via mcp-remote. See INSTALL.md in the package.

    It is free, and there is nothing to buy or sign up for. On currency: FHFA's index is repeat-sales, so new construction only enters after a home trades twice, and the series runs a quarter or two behind by construction. That matters little for a 10-to-30-year hold and more for a short one. Rent figures come from Zillow ZORI plus Census ACS.

    No, and it will refuse rather than guess. Price and rent figures are ZIP-level: good for a market, poor for one unit. There is no automated valuation and no listing lookup. Ask it for the market around an address, then supply the actual price or rent you're dealing with.

    No. US residential only. The underlying datasets — FHFA HPI, Census ACS, Zillow ZORI — are all domestic. It also does not currently model rental-investment metrics such as cap rate, DSCR or cash-on-cash; asking for those triggers a capability request rather than an invented answer.

    New in Finance & Accounting