Story Generation Pipeline Skill

Features

This skill implements a complete story generation pipeline:

1. Continuous Episode Generation - Automatically generates next episode based on previous content
Graph Management - Storage and querying of character, scene, and hook relationships
Dual Confirmation Control - AI quality check followed by human confirmation
State Persistence - Supports pause, resume, and multiple parallel stories

Core Workflow

CODEBLOCK0

Usage

Start New Pipeline

CODEBLOCK1

Continue Pipeline

CODEBLOCK2

User Confirmation Actions

CODEBLOCK3

Dual Confirmation Control Mechanism

Layer 1: AI Quality Check

After each episode is generated, AI automatically checks:

Check Item	Description
Plot Coherence	Natural connection with previous episode
Character Consistency

Scoring Standard: 0-10 points, below 7 triggers retry (max 3 times)

Layer 2: Human Confirmation

After AI review passes, display preview and wait for user confirmation:

CODEBLOCK4

Graph Management

Graph Query

Before generating episode N, query episode N-1's graph:

CODEBLOCK5

Returns:

- Complete content of previous episode
Character list and their statuses
Unclosed hooks
Key scenes
Relationship network

Graph Storage

After user confirmation, store current episode's complete content:

CODEBLOCK6

Stores: Complete generation result of the episode (not split into elements)

State Management

State File: `data/pipeline_state.json`

CODEBLOCK7

Status Types

Status	Description
INLINECODE1	Currently generating
INLINECODE2

Script Description

pipeline.py - Main Control Loop

- Initialize pipeline
Coordinate modules
Handle user commands
Manage loop state

ai_reviewer.py - AI Quality Check

- Execute quality scoring
Generate review report
Determine pass/fail

episode_generator.py - Episode Generation

- Generate new episode based on graph context
Handle hook continuation and closure
Handle retry logic

graph_manager.py - Graph Management

- Graph query (call remote API)
Graph storage (call remote API)
Local cache management

API Description

Graph API

Query API:
CODEBLOCK8

Storage API:

{
  "action": "save",
  "pipeline_id": "pipeline_2026-03-05-001",
  "episode": 3,
  "content": "Complete generation content for episode 3..."
}

Episode Generation Logic

First Episode Generation

Based on user-provided theme and goals, generate:

- Main character settings
Initial scenes
Core conflict
Open hooks

Subsequent Episode Generation

Based on graph query results:

1. Read previous episode content and unclosed hooks
Continue main plot line
Handle hooks (continue/close/add)
Advance character growth arc
Adjust emotional curve

Finale Generation

When target episodes reached or user requests end:

- Close all remaining hooks
Complete character growth arcs
Generate conclusive ending

Important Notes

1. Retry Mechanism - Max 3 retries when AI review fails
Pause/Resume - Can resume via pipeline_id after pause
Multiple Pipelines - Supports running multiple different-themed pipelines simultaneously
Graph Consistency - Ensure correct character and hook relationships
Hook Management - Track creation and closure status of each hook

Complete Workflow

Step 1: Create Pipeline

CODEBLOCK10

Step 2: Generate Episode

AI calls start_generation(pipeline_id) to get generation prompt, then generates episode content based on the prompt.

Step 3: Submit AI Review

AI calls submit_episode(pipeline_id, episode, content) to submit generated content, then executes AI review.

Step 4: Process Review Result

AI calls process_ai_review(pipeline_id, episode, ai_result, content) to process review results.

If review fails (score < 7), automatically retry (max 3 times).

Step 5: Wait User Confirmation

After review passes, display preview and wait for user confirmation:

CODEBLOCK11

Step 6: Process User Confirmation

After user confirms, AI calls user_confirm(pipeline_id, action, note):

- approve: Store graph, prepare next episode
modify: Regenerate based on feedback
pause: Pause pipeline
end: End pipeline

Step 7: Loop Generation

Repeat steps 2-6 until target episodes reached or user ends.

API Reference

Create Pipeline

CODEBLOCK12

Start Generation

CODEBLOCK13

Submit Episode

CODEBLOCK14

Process AI Review

CODEBLOCK15

User Confirm

CODEBLOCK16

Get Status

CODEBLOCK17

List Pipelines

CODEBLOCK18

Resume Pipeline

resume_pipeline(pipeline_id: str)
# Returns: Resume result

Example Dialogue

CODEBLOCK20

Important Notes

1. Retry Mechanism: Max 3 retries when AI review fails
State Persistence: All states saved in INLINECODE11
Graph Storage: Stored via remote API, requires network connection
Pause/Resume: Can resume via resume_pipeline after pause
Multiple Pipelines: Supports running multiple different-themed pipelines simultaneously

故事生成管线技能

功能特性

本技能实现完整的故事生成管线：

1. 连续剧集生成 - 基于前文内容自动生成下一集
图谱管理 - 存储和查询角色、场景及悬念关系
双重确认控制 - AI质量检查后由人工确认
状态持久化 - 支持暂停、恢复及多故事并行

核心工作流

┌──────────────┐ ┌──────────────┐ ┌──────────────┐ ┌──────────────────┐
│ 生成第N集 │ -> │ AI审核 │ -> │ 图谱存储 │ -> │ 等待人工确认 │
└──────────────┘ └──────────────┘ └──────────────┘ └──────────────────┘
│ │ │ │
▼ ▼ ▼ ▼
查询图谱通过/重试存储关系继续/修改/结束

使用方法

启动新管线

用户：启动新故事管线，主题：中国女孩逆袭，目标集数：10

继续管线

用户：继续故事管线 [pipeline_id]

用户确认操作

用户：批准，继续下一集
用户：修改：[具体反馈]
用户：暂停
用户：结束

双重确认控制机制

第一层：AI质量检查

每集生成后，AI自动检查：

检查项	描述
剧情连贯性	与上一集自然衔接
角色一致性

评分标准： 0-10分，低于7分触发重试（最多3次）

第二层：人工确认

AI审核通过后，显示预览并等待用户确认：

📋 第N集预览
━━━━━━━━━━━━━━━━━━━━━━━━
【剧情概要】
...

✅ AI评分：8.5/10
✅ 悬念状态：新增1个，闭合1个

请选择：
1️⃣ 批准，继续下一集
2️⃣ 需要修改（请说明）
3️⃣ 暂停
4️⃣ 结束
━━━━━━━━━━━━━━━━━━━━━━━━

图谱管理

图谱查询

生成第N集前，查询第N-1集的图谱：

python
graphmanager.querygraph(pipeline_id, episode=N-1)

- 上一集完整内容
角色列表及其状态
未闭合的悬念
关键场景
关系网络

图谱存储

用户确认后，存储当前集的完整内容：

python
graphmanager.savegraph(pipeline_id, episode=N, content)

存储：该集的完整生成结果（不拆分为元素）

状态管理

状态文件：data/pipeline_state.json

json
{
pipelines: {
pipeline_2026-03-05-001: {
theme: 中国女孩逆袭,
target_episodes: 10,
current_episode: 3,
status: waitinguserconfirm,
created_at: 2026-03-05T15:00:00,
updated_at: 2026-03-05T15:30:00,
ai_review: {
score: 8.5,
checks: {...}
},
last_output: {
episode: 3,
summary: ...,
content: 完整内容...
}
}
}
}

状态类型

状态	描述
generating	正在生成中
ai_reviewing

脚本说明

pipeline.py - 主控制循环

- 初始化管线
协调各模块
处理用户命令
管理循环状态

ai_reviewer.py - AI质量检查

- 执行质量评分
生成审核报告
判定通过/失败

episode_generator.py - 剧集生成

- 基于图谱上下文生成新剧集
处理悬念延续和闭合
处理重试逻辑

graph_manager.py - 图谱管理

- 图谱查询（调用远程API）
图谱存储（调用远程API）
本地缓存管理

API说明

图谱API

查询API：
json
{
action: query,
pipelineid: pipeline2026-03-05-001,
episode: 2
}

存储API：
json
{
action: save,
pipelineid: pipeline2026-03-05-001,
episode: 3,
content: 第3集完整生成内容...
}

剧集生成逻辑

首集生成

基于用户提供的主题和目标，生成：

- 主角设定
初始场景
核心冲突
开放悬念

后续剧集生成

基于图谱查询结果：

1. 读取上一集内容和未闭合悬念
延续主线剧情
处理悬念（延续/闭合/新增）
推进角色成长弧线
调整情感曲线

结局生成

当达到目标集数或用户要求结束时：

- 闭合所有剩余悬念
完成角色成长弧线
生成总结性结局

重要说明

1. 重试机制 - AI审核失败时最多重试3次
暂停/恢复 - 暂停后可通过pipeline_id恢复
多管线并行 - 支持同时运行多个不同主题的管线
图谱一致性 - 确保角色和悬念关系正确
悬念管理 - 追踪每个悬念的创建和闭合状态

完整工作流程

步骤1：创建管线

用户：启动新管线，主题：修仙少年，目标20集

AI：好的，正在创建管线 pipeline_20260305160000
主题：修仙少年
目标：20集
风格：写实电影感

状态：已初始化，准备生成第1集

步骤2：生成剧集

AI调用 startgeneration(pipelineid) 获取生成提示，然后根据提示生成剧集内容。

步骤3：提交AI审核

AI调用 submitepisode(pipelineid, episode, content) 提交生成内容，然后执行AI审核。

步骤4：处理审核结果

AI调用 processaireview(pipelineid, episode, airesult, content) 处理审核结果。

如果审核未通过（评分<7），自动重试（最多3次）。

步骤5：等待用户确认

审核通过后，显示预览并等待用户确认：

📋 第1集预览
━━━━━━━━━━━━━━━━━━━━━━━━
【修仙之路开启】

少年李云在山中发现一块神秘玉佩，
从此踏上修仙之路...

✅ AI评分：9.0/10
✅ 新增悬念：H-001 神秘玉佩的来历

━━━━━━━━━━━━━━━━━━━━━━━━

请选择：
1️⃣ 批准，继续下一集
2️⃣ 需要修改（请说明）
3️⃣ 暂停
4️⃣ 结束

步骤6：处理用户确认

用户确认后，AI调用 userconfirm(pipelineid, action, note)：

- approve：存储图谱，准备下一集
modify：根据反馈重新生成
pause：暂停管线
end：结束管线

步骤7：循环生成

重复步骤2-6，直到达到目标集数或用户结束。

API参考

创建管线

python createpipeline(theme: str, targetepisodes: int, style: str = realistic cinematic)

返回：{success: True, pipeline_id: ..., message: ...}

开始生成

python startgeneration(pipelineid: str)

返回：{success: True, episode: N, prompt: 生成提示}

提交剧集

python submitepisode(pipelineid: str, episode: int, content: str)

返回：{success: True, review_prompt: 审核提示}

处理AI审核

python processaireview(pipelineid: str, episode: int, airesult: str, content: str)

story-pipeline故事生成管线

story-pipeline

Story Generation Pipeline Skill

Features

Core Workflow

Usage

Start New Pipeline

Continue Pipeline

User Confirmation Actions

Dual Confirmation Control Mechanism

Layer 1: AI Quality Check

Layer 2: Human Confirmation

Graph Management

Graph Query

Graph Storage

State Management

State File: data/pipeline_state.json

Status Types

Script Description

pipeline.py - Main Control Loop

ai_reviewer.py - AI Quality Check

episode_generator.py - Episode Generation

graph_manager.py - Graph Management

API Description

Graph API

Episode Generation Logic

First Episode Generation

Subsequent Episode Generation

Finale Generation

Important Notes

Complete Workflow

Step 1: Create Pipeline

Step 2: Generate Episode

Step 3: Submit AI Review

Step 4: Process Review Result

Step 5: Wait User Confirmation

Step 6: Process User Confirmation

Step 7: Loop Generation

API Reference

Create Pipeline

Start Generation

Submit Episode

Process AI Review

User Confirm

Get Status

List Pipelines

Resume Pipeline

Example Dialogue

Important Notes

故事生成管线技能

功能特性

核心工作流

使用方法

启动新管线

继续管线

用户确认操作

双重确认控制机制

第一层：AI质量检查

第二层：人工确认

图谱管理

图谱查询

图谱存储

状态管理

状态文件：data/pipeline_state.json

状态类型

脚本说明

pipeline.py - 主控制循环

ai_reviewer.py - AI质量检查

episode_generator.py - 剧集生成

graph_manager.py - 图谱管理

API说明

图谱API

剧集生成逻辑

首集生成

后续剧集生成

结局生成

重要说明

完整工作流程

步骤1：创建管线

步骤2：生成剧集

State File: `data/pipeline_state.json`