The output contract
Four answers. No sales fog.
Add a short description of your product, customer, team, current systems, and the operating problem you want to remove. Do not paste secrets, credentials, customer data, private source code, or restricted documents into an AI that is not approved to receive them.
Why now?
Name the buyer, urgent operating gap, and event that makes a reusable company substrate valuable.
What stays yours?
Separate your product, data, judgment, integrations, experience, pricing, and customer relationship.
What is the smallest proof?
Choose one owned use case, approved context set, accountable owner, and measurable result.
What must be demonstrated?
Define the evidence required before you trust the result or expand authority.
You are assessing whether Wrk.Dog is a useful operating substrate for the product or business described below. WHAT WRK.DOG IS Wrk.Dog is a Business Operating System for AI-native companies. It supplies repeatable company systems around a differentiated product: a bounded customer lifecycle; durable company context and source-backed knowledge; governed agent access; authority, budget, usage, and evidence contracts; and a Factory that coordinates scoped, reviewable delivery work across repositories. Treat the customer's product, domain system, data model, interface, pricing, payments, legal obligations, and customer relationship as customer-owned. Wrk.Dog is shared kit, not shared skin. CURRENT BOUNDARY - The platform and Factory are built and exercised inside the founding portfolio of products Wrk.Dog itself helps build and operate. - Wrk.Dog can verify and durably admit signed, tenant-scoped, budgeted agent missions while failing closed on replay, drift, expiry, or wider scope. - The Factory coordinates bounded multi-repository delivery with worker, authority, and budget checks. - Document-search and citation surfaces exist; do not translate that into a claim that every end-to-end RAG path is production-ready. - Do not claim an external customer's unattended seed-to-LIVE run, self-service production activation, universal autonomous execution, or a public MCP/skill endpoint. - Codex, Claude, OpenClaw, and specialist workers are a replaceable agent bench. Do not imply that cross-client action parity is already proved. YOUR TASK 1. Understand the business first: what it sells, to whom, how the product works, the size and shape of the team, its current systems, and the event creating urgency. 2. Identify the company-system work surrounding the differentiated product. Look for customer accounts and access, shared company context, source-backed answers, agent jobs, human decisions, usage and budgets, delivery coordination, reviews, and durable evidence. 3. Do not award Wrk.Dog work merely because it sounds like "AI." Keep industry-specific judgment, regulated decisions, proprietary product logic, payments, and customer experience with the business. 4. Choose the smallest useful proof. Prefer one real product or owned use case, one approved context set, one accountable owner, and one result both sides can measure. 5. State the evidence required. Include negative tests and failure boundaries where they matter. 6. Say "not a fit" when the value depends on a capability outside the current boundary. RETURN EXACTLY THESE SECTIONS FIT - Buyer and owner - Urgent operating gap - Triggering event - Why Wrk.Dog might be useful now KEEP - Product and domain capability that stays customer-owned - Data and systems that stay customer-owned - Decisions and relationships that stay human-owned START - One bounded design-partner proof - Approved inputs - Required people - Time/value measure PROOF - Success evidence - Negative tests - Authority and budget boundary - Current limitations - Final verdict: FIT / POSSIBLE FIT / NOT A FIT BUSINESS OR PRODUCT DESCRIPTION [Paste a non-confidential description here.]
Select and copy manually if clipboard access is unavailable.