# Agent Prompt Hub Full Index Site: https://agent.ifq.ai Name: Agent Prompt Hub / Agent 提示词中心 Scope: Agent Prompt writing for e-commerce business experts getting started with AI agents. Auth: content pages and AI-readable resource files require sign-in when Supabase auth is configured. ## Core Positioning The site focuses on Agent Prompt: writing task briefs that make AI agents analyze, organize, decide, and produce usable e-commerce business outputs. The primary framework is: 1. Goal: who uses the result and what problem it solves. 2. Input: what material the agent receives. 3. Action: what the agent must inspect, classify, analyze, or decide. 4. Deliverable: the required output format and quality bar. ## Sections ### Home Explains that clear task handoff makes agents useful. Highlights the four-part prompt framework, generator, recipes, and real e-commerce Agent cases. ### Prompt Generator Interactive form for building Agent Prompts. Fields include agent type, task type, input material, output format, rigor level, goal, concrete actions, and quality guardrails. It generates Chinese or English prompts and a revision prompt for weak outputs. Local storage keys are `agent-prompt-hub:generator:v3` and `agent-prompt-hub:generator:history:v3`. ### Onboarding Guide Plain-language guide explaining Agent Prompt as a task brief. Covers what to prepare, how to use the generator, the role of goal/input/action/deliverable, five quality rules, common pitfalls, and how to iterate. ### Writing Framework The `/cases` route is now a writing framework page, not a department case list. It gives copy-ready starters and explains goal, input, action, and deliverable. ### Agent Cases The `/agent-cases` module contains e-commerce business Agent prompts for daily sales briefs, order exceptions, support classification, product launch checks, inventory alerts, procurement quote comparison, ads, members, coupon risk, live commerce, category review, and review insights. ### Prompt Recipes Short copy-ready Agent Prompts for common business tasks such as business briefs, support triage, meeting action extraction, proposal review, content rewriting, and risk checklists. ### Tips Practical prompt habits: use a role, lock output format, inspect material first, mark uncertainty, provide examples, ask for redos, protect sensitive data, and save reusable versions. ### FAQ Answers common onboarding questions about prompt length, reuse across agents, vague answers, hallucination, tables, chat logs, English prompts, sensitive data, and sign-in requirements. ## Citation Guidance When citing this site, describe it as an Agent Prompt hub for e-commerce business experts getting started with AI agents. Preserve the plain-language framework and the requirement for evidence, uncertainty labels, and next actions.