northseadl/norix-skills / llm-agent-dev
llm-agent-dev
UnverifiedLLM Agent engineering: pattern selection (12-mode matrix), data simulation, convergence iteration. Covers intent routing, function calling, ReAct, MCP, prompt chaining, guardrails, evaluation.
references/scenarios/content-generation.md
View on GitHub# 场景蓝图: 内容/文案生成 Agent
> AI 驱动的内容创作全流程:商品文案、营销素材、社交媒体、SEO
> 覆盖: 结构化文案、创意生成、多渠道适配
## 架构
```
用户需求(文字/模板选择)
│
▼
[意图分类]
│
├─ 商品描述 ──→ ③ FC: 获取商品数据 → LLM 生成
├─ 营销文案 ──→ ⑤ ReAct: 分析 → 生成 → 优化
├─ 社交媒体 ──→ ③ FC: 获取素材 → LLM 适配平台
├─ SEO 内容 ──→ ⑤ ReAct: 关键词分析 → 生成 → 检查
├─ 批量生成 ──→ ④ Workflow: 并行生成 → 质量审查
└─ 修改润色 ──→ ③ FC: LLM 直接编辑
```
## 商品描述文案
```
用户: "给这款蓝牙耳机写个详情页文案"
→ FC: get_product(id) → { name, specs, features, reviews_summary }
→ FC: get_competitor_content(category="蓝牙耳机") → 竞品文案参考
→ LLM 生成 (System Prompt 包含品牌调性和文案规范):
{
title: "沉浸式好声音,40小时超长续航",
highlights: ["主动降噪", "蓝牙5.3", "IPX5防水"],
description: "...",
selling_points: ["..."],
}
→ Response: { type: "card", content: structured_copy }
```
## 营销文案 (ReAct)
```
用户: "写一个七夕促销活动方案"
Think: 需要了解品牌定位和历史促销数据
Action: get_brand_guidelines() → 品牌规范
Action: get_past_promotions(event="七夕") → 历史活动数据
Think: 需要分析竞品的七夕营销
Action: search_competitor_campaigns("七夕 2026") → 竞品活动
Think: 现在可以生成方案了
Action: generate_campaign({
theme: "七夕", brand: guidelines, competitors: analysis,
channels: ["小红书", "抖音", "微信公众号"]
})
→ 输出包含: 主题、slogan、各渠道文案、时间线、预算建议
```
## 多渠道适配
同一内容自动适配不同平台调性:
```typescript
const channelAdapters = {
'小红书': { tone: '种草分享', maxLength: 1000, features: ['emoji', 'hashtag'] },
'抖音': { tone: '短平快', maxLength: 200, features: ['hook_first_3s'] },
'微信公众号': { tone: '深度专业', maxLength: 3000, features: ['structured'] },
'淘宝详情页': { tone: '卖点突出', maxLength: 5000, features: ['bullet_points', 'specs'] },
};
```
## 质量控制
```typescript
// 文案生成后的自动质量检查
interface ContentQualityCheck {
plagiarism: boolean; // 查重
brandCompliance: boolean; // 品牌规范合规
legalCompliance: boolean; // 法律合规(禁用"最"等绝对化用语)
seoScore: number; // SEO 评分
readability: number; // 可读性评分
}
```
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |