Yao Demand Skill
Businessyao-demand-skillYao Demand Skill is an Agent Skill for business published in the yaojingang/yao-open-skills repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.
Evaluate product demand from product links, product descriptions, PRDs, websites, app-store pages, white papers, sales decks, screenshots, or funding materials using the demand triangle model: lack, target object, and consumer ability. Use when asked for demand assessment, pre-investment product review, growth diagnosis, positioning validation, competitor-backed demand evidence, or a multi-format demand report. Do not use for pure market sizing, generic business-model design, UX-only review, legal/financial advice, or ad-copy ideation without demand evidence.
How to use this skill
Bring this guide into your coding agent with a prompt tailored to the tool you use.
- Open your project in Codex.
- Copy and paste the prompt below.
- Review files and risks before approving any action.
Inspect the Agent Skill at https://github.com/yaojingang/yao-open-skills/blob/HEAD/skills/yao-demand-skill/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/yao-demand-skill, but wait for my approval before writing files or running scripts.
Copying this prompt does not install or run the skill. Review third-party files before use.
Yao Demand Skill
Evidence-backed demand assessment for products, services, apps, SaaS, AI tools, consumer goods, education products, and early-stage ventures.
Use This Skill For
- assessing whether a product has a solid demand foundation before building, investing, launching, or scaling
- diagnosing weak conversion, weak retention, vague positioning, pricing friction, trust friction, or adoption barriers
- comparing direct competitors, indirect substitutes, current user workarounds, and the option of not buying
- producing a visual demand diagnosis report with citations, scores, red flags, 10+ chart modules, experiments, forecasts, and four final formats
Do Not Route Here
- pure TAM/SAM/SOM market sizing without product-demand diagnosis
- generic monetization or business-model option design; use a business-model skill instead
- UX heuristic review without demand, JTBD, or adoption evidence
- legal, financial, medical, or investment advice as a final decision
- manipulative marketing designed to shame, scare, addict, or exploit vulnerable users
Workflow
- Confirm the product input. Accept a URL, text description, PRD, website copy, docs, screenshot, app-store page, sales material, or funding deck. Ask only one concise question if no product substance is available.
- Build the product canvas: product definition, user, scenario, features, price, promise, business model, market, source list, and unresolved assumptions.
- Plan evidence. Prioritize official sources, third-party validation, user feedback, competitor/substitute evidence, and time-sensitive market or regulatory facts.
- Research only evidence that can support or challenge demand. Current product, price, competitor, market, legal, or regulatory facts must be verified with sources and dates.
- Segment users by JTBD, trigger scenario, buying role, current alternatives, and adoption blockers.
- Analyze the three demand triangle dimensions:
lack,target_object, andconsumer_ability. Include evidence, counter-evidence, assumptions, and improvement paths. - Score each dimension from
0to10, then calculate total score with the geometric short-board formula and confidence adjustment. - Produce visual diagnostics: at least 10 chart modules, each with one or two insight sentences, one recommendation, confidence, and evidence or assumption binding.
- Produce recommendations, forecast scenarios, and a final 30/60/90 day action plan: positioning, product, pricing, onboarding, trust, channel, and validation experiments.
- Run QA: citation coverage, time consistency, evidence diversity, at least three counter-signals, score explainability, chart completeness, ethics, and layout readiness.
- Write a structured report JSON, then use
scripts/render_report.pyto create Markdown, HTML, Word, and PDF outputs.
Output Contract
- Always produce the final report in four formats:
.md,.html,.docx, and.pdf. - Use one canonical
report.jsonas the rendering source when possible, so the four outputs remain consistent. - HTML must include a top follow menu bar that stays pinned while the page scrolls, with quiet anchor navigation.
- Formal reports must follow a
summary -> visual diagnostics -> deep analysis -> final planstructure. - Formal reports must include at least 10 chart modules. HTML/PDF render them as inline SVG. Markdown and Word must include chart-equivalent tables or images.
- All report backgrounds are pure white. Borrow Kami's editorial hierarchy, ink-blue accent, table discipline, typography, spacing, and production checks, but override Kami's parchment background.
- Every key factual claim must either cite a source ID or be labeled as an assumption.
- Every score must include evidence, reasoning, uncertainty, and a concrete improvement path.
- Forecasts must be scenario-based and labeled with assumptions and confidence. Do not present uncertain adoption outcomes as deterministic predictions.
Reference Map
- Read
references/workflow.mdbefore starting an assessment. - Read
references/evidence-policy.mdbefore using sources, citations, or current facts. - Read
references/triangle-model.mdbefore scoring. - Read
references/report-contract.mdbefore writing the report JSON or final narrative. - Read
references/kami-white-report-layout.mdbefore rendering the four output formats. - Use
templates/report.schema.jsonas the report JSON target. - Use
scripts/score_triangle.pyto calculate or verify weighted scores. - Use
scripts/validate_report.pybefore rendering. - Use
scripts/render_report.pyto generate Markdown, HTML, Word, and PDF.