Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/qiuxchao/content-agent/claude-mdgit clone --depth 1 https://github.com/qiuxchao/content-agentWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/qiuxchao/content-agent/claude-md)<a href="https://agentmods.dev/instructions/qiuxchao/content-agent/claude-md"><img src="https://agentmods.dev/badge/instructions/qiuxchao/content-agent/claude-md.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00756 | $0.00756 |
| Opus 5 | $0.00378 | $0.00378 |
| Sonnet 5 | $0.00151 | $0.00151 |
| Haiku 4.5 | $0.00076 | $0.00076 |
Grade A, and why
content-agent CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
Content Agent
AI 内容生产 Agent,基于 LangGraph + LangChain,输入主题自动生成多平台文章(微信公众号 / 小红书 / 知乎)。
技术栈
- Agent: Python 3.11+, LangGraph, LangChain
- LLM: OpenAI 兼容(DeepSeek/Kimi/通义等)+ Anthropic Claude
- 搜索: Tavily
- 向量库: Chroma(本地持久化
data/vectorstore/) - 数据库: SQLite(
data/content-agent.db,WAL 模式) - API: FastAPI + SSE 流式输出
- 前端: Next.js + React 19 + Ant Design + Tailwind CSS
- 包管理: Python 用
uv,前端用bun
启动方式
# API 服务
uv run uvicorn api.server:app --reload --port 8917
# 前端(另一个终端)
cd web && bun dev
# 访问 http://localhost:3917
项目结构
agent/
graph.py # LangGraph 主图定义
state.py # AgentState 共享状态
llm.py # LLM 工厂(OpenAI/Anthropic)
config.py # 配置读取(env > SQLite > default)
db.py # SQLite 数据库
memory.py # Chroma 向量库(RAG)
nodes/ # 图节点
pre_researcher → planner → researcher → writer → critic
↑ (retry ≤2, score <7) ↓ (pass)
└── increment_retry ←───┘
image_fetcher → save_memory → END
prompts/templates.py # 平台 Prompt 模板 + 方向预设
publish/ # 微信发布(MD→HTML、API、封面)
tools/ # 搜索、生图、截图、Unsplash
api/server.py # FastAPI 路由
web/ # Next.js 前端
配图系统
Writer 在初稿中插入占位符,ImageFetcher 节点替换为真实图片。两种占位符:
[IMAGE: 详细英文绘图提示词]— AI 生图或 Unsplash 搜图[SCREENSHOT: url, 中文描述]— Playwright 截取网页
通过 IMAGE_PROVIDER 配置控制模式:
prompt— 仅生成提示词,用户手动上传screenshot— 纯截图模式mixed— 混合模式(Writer 自行选择截图或生图)unsplash/openai/gemini等 — AI 生图
开发约定
- 配置优先级:环境变量 > SQLite settings 表 > 默认值
- LLM 实例通过
@lru_cache缓存,配置变更需重启 - 前端设置修改写入 SQLite,无需重启即生效(热更新)
- 图片保存到
data/images/,通过/api/images/{filename}访问 - SSE 流式事件格式:
{node, data, active},active字段标记当前执行节点 - 中文正文,英文专有名词保留原文
- 修改代码时,要考虑是否需要同步全局注释/文档
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 6d ago First seen · 71 lines · 756 tokens per session scan A 59ca2ad9857b
content-agent CLAUDE.md is an instructions file published in the GitHub repository qiuxchao/content-agent (24 stars, last pushed 5mo ago), licensed MIT. It adds 756 tokens to every session, about $0.0038 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
LangBot AGENTS.md
AGENTS.md instructions for langbot-app/LangBot, covering agents.md, quick facts, essential commands, where to look and cross-repo sdk work.
she-love-me AGENTS.md
Instructions for 863401402/she-love-me, covering agents.md, repository focus and codex guidance.
she-love-me CLAUDE.md
Instructions for 863401402/she-love-me, covering claude.md and 唯一工作流.
she-love-me copilot-instructions.md
Instructions for 863401402/she-love-me: When the user asks to import or analyze WeChat/QQ chat history, read and follow .agents/skills/she-love-me/SKILL.md from the repository root.
she-love-me GEMINI.md
Instructions for 863401402/she-love-me: For WeChat or QQ chat import and relationship analysis, use .agents/skills/she-love-me/SKILL.md as the authoritative workflow. Read its referenced files when routed, especially references/data-sources.md for end-to-end exporter installation and operation.
LangBot CLAUDE.md
Claude Code instructions for langbot-app/LangBot, a project described as: Production-grade platform for building agentic IM bots - 生产级多平台智能机器人开发平台/ Agent、知识库编排、插件系统 / Bots for Discord / Slack / LINE / Telegram / WeChat(企业微信, 企微智能机器人, 公众号) / 飞书 / 钉钉 / QQ / Matrix e.g. Integrated with ChatGPT(GPT), DeepSeek, Dify, n8n…