SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillslark-workflow-content-pipeline

lark-workflow-content-pipeline

Office & productivity

View repository

Updated 10-16 06:23

About this skill

Content Creation Pipeline Workflow

CRITICAL — Before starting, you MUST first use the Read tool to read ../lark-shared/SKILL.md, which contains authentication and permission handling

Applicable Scenarios

  • "This is a content idea" / "Mark as content idea"
  • "Help me evaluate this topic" / "Topic evaluation"
  • "Research this topic" / "Topic research"
  • "Generate a content proposal" / "Write an outline"
  • "What's in my idea library?" / "Idea list"

Prerequisites

Only supports user identity. Ensure authorization before execution:

lark-cli auth login --domain im,task,wiki,base,docs

Workflow Overview

创意触发：
  用户标记消息/直接描述 ──► AI 读取上下文

话题研究：
  ├── WebSearch ──────────► 搜索最新动态
  ├── doc +search ─────────► 检查知识库重复选题
  └── base +record-list ───► 检查创意库重复

评估与方案：
  AI 评估价值 ──► 生成方案（标题/大纲/开头）

任务创建：
  task +create ──► 创建飞书任务跟踪
  base +record-upsert ──► 存入创意库

通知：
  im +messages-send ──► 回复完成通知

Data Initialization

Base Table Structure

Table 1: Idea Library (idea_library)

Field Name Type Description
Idea Title Text Short title of the idea
Source Single select Group chat/Conversation/Manual/Other
Source Details Text Source group chat name/conversation context
Original Description Text Original description of the idea
Topic Domain Single select Technology/Business/AI/Product/Other
Evaluation Score Number 1-10 points (AI evaluation + user adjustment)
Status Single select Inspiration/Researching/Proposal/In Progress/Completed/Abandoned
Duplicate Check Single select Passed/Suspected duplicate/Duplicate
Duplicate Source Text If duplicate, points to existing content
Associated Task ID Text Feishu task task_id
Proposal Document Link URL Link to the generated proposal document
Creation Time Date Time added to library
Update Time Date Last update time

Table 2: Idea Research (idea_research)

Field Name Type Description
Associated Idea Relation Linked to "Idea Library"
Research Source Text Search source URL or description
Research Type Single select Latest Updates/Competitor Analysis/User Discussion/Data Support
Summary Text Summary of research content
Relevance Single select High/Medium/Low

Initialization Command

# 注意：+table-create 可能部分成功，需捕获错误后用 +field-create 补字段
# 注意：连续创建字段会触发限流（错误码 800004135），每次调用间隔至少 1 秒
lark-cli base +table-create --name "创意库" --fields '[
  {"name":"创意标题","type":"text"},
  {"name":"来源","type":"select","multiple":false,"options":[{"name":"群聊"},{"name":"对话"},{"name":"手动"},{"name":"其他"}]},
  {"name":"来源详情","type":"text"},
  {"name":"原始描述","type":"text"},
  {"name":"话题领域","type":"select","multiple":false,"options":[{"name":"科技"},{"name":"商业"},{"name":"AI"},{"name":"产品"},{"name":"其他"}]},
  {"name":"评估分数","type":"number"},
  {"name":"状态","type":"select","multiple":false,"options":[{"name":"灵感"},{"name":"研究中"},{"name":"方案"},{"name":"执行中"},{"name":"已完成"},{"name":"已放弃"}]},
  {"name":"重复检查","type":"select","multiple":false,"options":[{"name":"通过"},{"name":"疑似重复"},{"name":"重复"}]},
  {"name":"重复来源","type":"text"},
  {"name":"关联任务ID","type":"text"},
  {"name":"方案文档链接","type":"link"},
  {"name":"创建时间","type":"datetime"},
  {"name":"更新时间","type":"datetime"}
]'
lark-cli base +table-create --name "创意研究" --fields '[
  {"name":"研究来源","type":"text"},
  {"name":"研究类型","type":"select","multiple":false,"options":[{"name":"最新动态"},{"name":"竞品分析"},{"name":"用户讨论"},{"name":"数据支持"}]},
  {"name":"摘要","type":"text"},
  {"name":"相关性","type":"select","multiple":false,"options":[{"name":"高"},{"name":"中"},{"name":"低"}]}
]'
# 注：关联创意字段需单独用 +field-create 创建（type="lookup" 需要关联目标表 ID）

Rate Limit Handling: If +table-create partially fails due to rate limiting, the table will still be created successfully. In this case:

  1. Use base +table-list to get the table_id
  2. Use base +field-create --table-id <id> --json '{"name":"字段名","type":"text"}' to add fields one by one, with an interval of 1-2 seconds each time

Mode 1: Idea Triggering and Processing

Method A: Marking from Feishu group chat

The user marks a message in the group chat as "content idea":

# 获取群聊消息及上下文
lark-cli im +chat-messages-list --chat-id "<chat_id>" --count 20

The AI reads the message context to understand the full background of the idea.

Method B: Direct description

The user directly describes the idea in the conversation:

  • "I think we can write an article about AI Agent implementation in enterprises"
  • "This topic is good, help me turn it into a topic selection"

Processing flow

创意触发 ──► AI 理解背景 ──► 提炼核心创意
                                  │
                                  ▼
                          进入研究流程

Mode 2: Topic Research

Step 1: Search for the latest updates

Use AI capabilities to search for the latest updates on related topics:

  • Search engine search: latest articles, discussions, trends
  • Technical communities: related technical progress

Step 2: Check for duplicate topic selections

Check the idea library:

lark-cli base +record-list --table-id "<idea_table_id>" --filter '...话题领域 = "<领域>"'

Check the knowledge base (linked with the knowledge base Skill):

lark-cli docs +search --query "<话题关键词>"

Step 3: Evaluate topic popularity

The AI evaluates based on search results:

  • The recent discussion popularity of the topic
  • Whether it is time-sensitive (hot events)
  • Whether competitors have already covered similar topics
  • The level of interest of the target audience

Step 4: Record research results

lark-cli base +record-upsert --table-id "<research_table_id>" --json '{"字段名":"值"}'

Mode 3: Proposal Generation

The AI evaluates the value of the idea

Evaluation dimensions (each scored 1-10):

Dimension Weight Description
Timeliness 25% Whether it is a current hot topic
Differentiation 25% The degree of difference from existing content
Audience value 20% The value to the target audience
Executability 15% The degree to which the idea can be executed
Extensibility 15% Whether it can be made into a series

Generate content proposal

## 内容方案

### 基本信息
- 标题建议：{3个备选标题}
- 内容类型：{文章/视频/播客/系列}
- 目标字数：{建议字数}
- 预计耗时：{预估}

### 开头建议（Hook）
> {一个能抓住读者注意力的开头段落}

### 详细大纲
1. {第一章标题}
   - {要点1}
   - {要点2}
2. {第二章标题}
   - {要点1}
   - {要点2}
...

### 关键数据/引用
- {支撑论点的数据和引用}

### 发布建议
- 推荐平台：{平台列表}
- 最佳发布时间：{时间建议}
- 标签建议：{标签列表}

### 评估分数：{总分}/10
- 时效性：{分} | 差异化：{分} | 受众价值：{分}
- 可执行性：{分} | 延展性：{分}

Save the proposal as a Feishu document

lark-cli docs +create --title "内容方案：{创意标题}" --markdown "<方案内容>"

Mode 4: Task Creation

After the user confirms the proposal, automatically create a Feishu task for tracking:

lark-cli task +create --summary "创作：{创意标题}" --description "内容方案已生成，详见文档链接" --due "<建议完成日>"

Update the idea library status:

lark-cli base +record-upsert --table-id "<idea_table_id>" --json '{
  "创意标题": "...",
  "状态": "执行中",
  "关联任务ID": "...",
  "方案文档链接": "..."
}'

Mode 5: Idea Library Management

View the idea list

lark-cli base +record-list --table-id "<idea_table_id>" --filter '...状态 != "已完成" AND 状态 != "已放弃"' --sort '评估分数 DESC'

Update idea status

  • "Mark this idea as completed" → update the status to "Completed"
  • "Abandon this idea" → update the status to "Abandoned"
  • "Adjust the evaluation score" → update the evaluation score

Permission table

Command Required scope
im +chat-messages-list im:message:readonly
im +messages-send im:message:send_as_bot
task +create task:task:write
docs +create docx:document:create
docs +search drive:drive:read
base +table-create bitable:app
base +record-list bitable:app:read
base +record-upsert bitable:app:write

Reference

  • lark-shared — Authentication, permissions (required reading)
  • lark-im — +chat-messages-list、+messages-send Detailed usage
  • lark-task — +create Detailed usage
  • lark-doc — +create、+search Detailed usage
  • lark-base — +table-create、+record-list、+record-upsert Detailed usage
  • lark-workflow-knowledge-base — Knowledge base linkage (duplicate check)