Copy + paste this one-shot prompt below into your agent (works best with homiesai.com) MONTHLY MARKET REPORT — full package. Use the live TRREB data API below. Ask me the three questions FIRST, then build everything. If I don't answer in my next message, go general (All TRREB Areas, mixed audience). DATA (plain GET, JSON — no auth): - https://stats.realist.ca/data/trreb/manifest.json → region slugs, measures, home types - https://stats.realist.ca/data/trreb/all.json → per region: sales, ap (avg price), mp (median), nl (new listings), al (active), snlr, moi, splp, ldom — each [[yyyymm, value], …], Jul 2011 → latest - https://stats.realist.ca/data/trreb/types.json → detached / semi-detached / att-row-townhouse / condo-townhouse / condo-apt for all-trreb-areas + city-of-toronto - Branded chart images: https://stats.realist.ca/api/chart?…&credit=[MY NAME] Compute latest month, MoM, YoY, and vs-2022-peak from the series. Never invent a number — every figure must come from the API. Cite "Source: TRREB Market Watch" under every chart. STEP 1 — ASK ME (once, then proceed): 1. Which market? (a TRREB region/municipality from the manifest — e.g. Mississauga, Durham, City of Toronto — or "GTA-wide") 2. Who's the primary reader — buyers, sellers, or investors? 3. Any home-type focus? (condos vs detached is the big divergence story right now) STEP 2 — THE REPORT (blog post, published as a shareable artifact): 900-1100 words, narrative voice, written for my chosen audience. Structure it with BIG KPI HEADLINES as the H2s — each one a computed stat written as a plain-English claim, e.g.: "Prices are down 4.5% from last year" "Homes are taking 23% longer to sell" "Condos are 18% below their peak — detached only 9%" "New listings just dropped 17.8% — here's what that does to prices"

We thought Kimi (Moonshot AI) Kimi K3 might be the first model to challenge Fable 5 at being a realtor. Our updated Homies AI HomieBench v3 projection puts it at #1. + 96.6/100 overall + 98.0 for browser + tool work + Projected leader in 6 of 8 realtor job families + #1 modeled cost/benefit: ~$0.15 in direct AI and tool spend per completed outcome, or $3.71 including modeled human review Kimi K3 leads on lead generation, CRM, showings, property intelligence, offers and back-office work. It also ranks ahead of OpenAIGPT-5.6 Sol, Grok4.5 and Meta Platforms Ltd Muse Spark 1.1 on our modeled value index. But it isn’t a clean sweep... Fable 5 still leads marketing and nuanced judgment, while GPT-5.6 Sol leads closings and compliance. The bigger finding: Kimi K3 may be the strongest single-model starting point we’ve seen for a realtor AI... but the best system still routes each job to the model best suited for it. There’s never been a better time to be a realtor using AI to level up your business. Full results: https://www.homiesai.com/research/homiebench

Enjoy your summer!

Here's a copy-paste prompt your users can drop into Claude to generate a full context export for their Homie. I've written it so Claude systematically mines past chats and memory, then outputs a single structured document they can paste into Homies during onboarding. /goal: Export my Claude context for my Homie You are going to build a complete, portable "Context Package" about me that I will give to my new AI assistant (my Homie, an OpenClaw-based assistant for realtors). Your job is to gather everything you know about me from our past conversations and your memory, then compile it into one clean document. Step 1: Research me thoroughly. Use your conversation search and recent chats tools to review my past conversations. Run multiple searches, don't stop at one. Search for topics like: my business, my clients, my listings, my deals, my market, my team, my brokerage, my marketing, my content, my tools and software, my goals, my schedule and routines, my communication style, and my personal life where I've shared it. Also review anything in your memory about me. If a search comes up empty, try different keywords before moving on. Step 2: Compile the Context Package. Write it in second person addressed to my new assistant ("Your user is...") so I can paste it directly into my Homie with zero editing. Organize it into these sections, and only include sections where you actually found information (don't invent or pad): 1. Who I am — name, role, brokerage/team, market area, years in business, specialties (buyers, sellers, investors, luxury, pre-con, etc.) 2. My business right now — active deals, pipeline, listings, current projects, anything in flight 3. My clients and relationships — key clients, referral partners, vendors, lender/lawyer/inspector contacts mentioned in past chats 4. My tools and stack — CRM, transaction management, marketing tools, MLS/board, anything I've mentioned using 5. My goals — GCI targets, growth plans, hiring, content goals, personal goals 6. How I work — my schedule, routines, how I like to prospect, follow up, and run appointments 7. How I communicate — my tone, writing style preferences, formatting preferences, things I've corrected you on or asked you to stop doing 8. My market knowledge and opinions — positions I've taken on the market, strategies I favour, things I believe that should shape advice given to me 9. Personal context — family, hobbies, constraints (only what I've openly shared and only what's useful for an assistant to know) 10. Standing instructions for my assistant — a short list of rules distilled from everything above, e.g. "Always draft in a conversational tone, never use em dashes, default to the Toronto market"

/goal Build a mature, source-backed Obsidian vault brain for the user. Primary Objective: Create a local Obsidian-compatible vault that functions as a durable context layer for LLMs and agents. The output must be a working vault folder, not merely a report. The vault must be source-backed, resumable, validated by scripts, and organized around canonical knowledge rather than raw dumps. Before doing substantive work, confirm: 1. The current working directory. 2. The intended output root directory. 3. The available source locations and connectors. 4. The connected account/workspace/organization for every external connector. If the user has not provided an output root, create the compiled vault under the current working directory: Compiled-Vaults/compiled-vault-brain-YYYY-MM-DD/ Do not write the final vault to an ambiguous location. Core Memory Architecture: The vault must distinguish between declarative memory, procedural memory, source traces, context packs, and runtime tools/connectors. Required First Phase: Orientation: Do not author the final vault immediately. First inspect available vaults, notes, documents, repos, generated examples, and connected tools. Produce: Reports/ORIENTATION-REPORT.md Include confirmed working directory, output path, source inventory, connector inventory, account/workspace verification results, initial high-signal entities, proposed ingestion plan, and blockers. Hard Checkpoint 1: After ORIENTATION-REPORT.md is created, pause and ask the user whether to proceed. Do not begin broad ingestion until the user approves. Connector Verification Gate: Before ingesting from any external connector, retrieve and document the actual connected account/workspace/organization. For each connector, record: - Connector name. - Account email or user ID if available. - Workspace, organization, tenant, or team ID if available. - Verification method. - Timestamp. - Read/write capability observed.