addon-langchain-llm

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Add-on: LangChain LLM

附加组件：LangChain LLM

Use this skill when an existing project needs LangChain primitives for chat, retrieval, or summarization.

当现有项目需要用于聊天、检索或摘要功能的LangChain基元时，可使用此技能。

Compatibility

兼容性

Works with

architect-python-uv-fastapi-sqlalchemy

architect-python-uv-batch

, and

architect-nextjs-bun-app

Can be combined with
```
addon-rag-ingestion-pipeline
```
.
Can be combined with
```
addon-langgraph-agent
```
when graph orchestration is required.
Can be combined with
```
addon-llm-judge-evals
```
; when used together, declare
```
langchain
```
in
```
config/skill_manifest.json
```
so the judge runner can resolve the backend without guessing.

适配

architect-python-uv-fastapi-sqlalchemy

、

architect-python-uv-batch

和

architect-nextjs-bun-app

技术栈。

可与
```
addon-rag-ingestion-pipeline
```
搭配使用。
需要图编排能力时，可与
```
addon-langgraph-agent
```
搭配使用。
可与
```
addon-llm-judge-evals
```
搭配使用；同时使用时，请在
```
config/skill_manifest.json
```
中声明
```
langchain
```
，以便评测运行器可以直接解析后端，无需猜测。

Inputs

输入参数

Collect:

```
LLM_PROVIDER
```
:
```
openai
```
|
```
anthropic
```
|
```
ollama
```
.
```
DEFAULT_MODEL
```
: provider model id.
```
ENABLE_STREAMING
```
:
```
yes
```
|
```
no
```
(default
```
yes
```
).
```
USE_RAG
```
:
```
yes
```
|
```
no
```
.
```
MAX_INPUT_TOKENS
```
: default
```
8000
```
.

需收集以下信息：

```
LLM_PROVIDER
```
：
```
openai
```
|
```
anthropic
```
|
```
ollama
```
。
```
DEFAULT_MODEL
```
：对应提供商的模型ID。
```
ENABLE_STREAMING
```
：
```
yes
```
|
```
no
```
（默认值为
```
yes
```
）。
```
USE_RAG
```
：
```
yes
```
|
```
no
```
。
```
MAX_INPUT_TOKENS
```
：默认值为
```
8000
```
。

Integration Workflow

集成工作流

Add dependencies:

Python:

bash

uv add langchain langchain-core langchain-community pydantic-settings tiktoken

Next.js:

bash

bun add langchain zod

Provider packages (as needed):

bash

uv add langchain-openai langchain-anthropic langchain-ollama
bun add @langchain/openai @langchain/anthropic @langchain/ollama

Add files by architecture:

Python API:

text

src/{{MODULE_NAME}}/llm/provider.py
src/{{MODULE_NAME}}/llm/chains.py
src/{{MODULE_NAME}}/api/routes/llm.py

Next.js:

text

src/lib/llm/langchain.ts
src/lib/llm/chains.ts
src/app/api/llm/chat/route.ts

Enforce typed request/response contracts:

Validate input lengths before chain invocation.
Return stable schema for streaming and non-streaming modes.

If
```
USE_RAG=yes
```
, compose retriever + prompt + model chain:

Keep retrieval source metadata in outputs.
Bound document count and token budget.

If
```
addon-llm-judge-evals
```
is also selected:

emit

config/skill_manifest.json

with

addon-langchain-llm

addons

declare

"judge_backends": ["langchain"]

capabilities

allow the judge runner to reuse
```
DEFAULT_MODEL
```
when
```
JUDGE_MODEL
```
is unset

添加依赖：

Python 环境：

bash

uv add langchain langchain-core langchain-community pydantic-settings tiktoken

Next.js 环境：

bash

bun add langchain zod

提供商对应包（按需安装）：

bash

uv add langchain-openai langchain-anthropic langchain-ollama
bun add @langchain/openai @langchain/anthropic @langchain/ollama

按架构添加对应文件：

Python API：

text

src/{{MODULE_NAME}}/llm/provider.py
src/{{MODULE_NAME}}/llm/chains.py
src/{{MODULE_NAME}}/api/routes/llm.py

Next.js：

text

src/lib/llm/langchain.ts
src/lib/llm/chains.ts
src/app/api/llm/chat/route.ts

强制使用带类型的请求/响应契约：

在调用链之前验证输入长度。
为流式和非流式模式返回稳定的schema。

若
```
USE_RAG=yes
```
，组合检索器 + 提示词 + 模型链：

在输出中保留检索源元数据。
限制文档数量和token预算。

若同时选择了
```
addon-llm-judge-evals
```
：

生成

config/skill_manifest.json

，在

addons

字段中加入

addon-langchain-llm

在

capabilities

字段中声明

"judge_backends": ["langchain"]

当
```
JUDGE_MODEL
```
未设置时，允许评测运行器复用
```
DEFAULT_MODEL
```

Required Template

所需模板

Chat response shape

聊天响应结构

json

{
  "outputText": "string",
  "model": "string",
  "provider": "string"
}

json

{
  "outputText": "string",
  "model": "string",
  "provider": "string"
}

Guardrails

防护规则

Documentation contract for generated code:
- Python: write module docstrings and docstrings for public classes, methods, and functions.
- Next.js/TypeScript: write JSDoc for exported components, hooks, utilities, and route handlers.
- Add concise rationale comments only for non-obvious logic, invariants, or safety constraints.
- Apply this contract even when using template snippets below; expand templates as needed.
Enforce provider/model allow-lists.
Add timeout and retry limits around provider calls.
Never log secrets or raw auth headers.
On streaming disconnect, stop upstream generation promptly.
If judge evals are enabled, keep the judge path on the same provider abstraction instead of bypassing it with ad hoc SDK calls.

生成代码的文档契约：
- Python：为模块、公开类、方法和函数编写文档字符串。
- Next.js/TypeScript：为导出的组件、钩子、工具函数和路由处理器编写JSDoc。
- 仅为非显见逻辑、不变量或安全约束添加简明的原理注释。
- 即使使用下方的模板片段也需遵守此契约，可根据需要扩展模板。
强制实施提供商/模型白名单。
为提供商调用添加超时和重试限制。
永远不要记录密钥或原始认证头。
流式连接断开时，及时停止上游生成。
若启用了评测功能，需将评测路径保持在同一提供商抽象层上，不要通过临时SDK调用绕过该层。

Validation Checklist

验证检查清单

Confirm generated code includes required docstrings/JSDoc and rationale comments for non-obvious logic.

bash

uv run ruff check . || true
uv run mypy src || true
bun run lint || true
rg -n "langchain|outputText|provider" src

Manual checks:
Typed chat route returns valid response.
Invalid payloads fail with controlled validation errors.

确认生成的代码包含所需的文档字符串/JSDoc，以及非显见逻辑的原理注释。

bash

uv run ruff check . || true
uv run mypy src || true
bun run lint || true
rg -n "langchain|outputText|provider" src

人工检查项：
带类型的聊天路由返回有效响应。
无效载荷会返回受控的验证错误。

Decision Justification Rule

决策论证规则

Every non-trivial decision must include a concrete justification.
Capture the alternatives considered and why they were rejected.
State tradeoffs and residual risks for the chosen option.
If justification is missing, treat the task as incomplete and surface it as a blocker.

每个非琐碎决策都必须包含具体的论证依据。
记录考虑过的替代方案以及拒绝的原因。
说明所选方案的权衡和残留风险。
若缺少论证，视为任务未完成，将其标记为阻塞项。