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/jackjin1997/redpress/claude-mdgit clone --depth 1 https://github.com/jackjin1997/redpressWhat 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 | $0.00939 | $0.00939 |
| Opus 5 | $0.00469 | $0.00469 |
| Sonnet 5 | $0.00188 | $0.00188 |
| Haiku 4.5 | $0.00094 | $0.00094 |
Grade A, and why
redpress 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 3d 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.
How it starts
The opening of the file, as written. The whole thing — 63 lines — stays where its author put it; the contents beside it link to each section on GitHub.
redpress
小红书的发布印刷机——草稿进,成品图文帖出。
项目定位与边界
redpress 管"怎么发",lab 管"发什么、效果如何"。
- 本仓库(公开):渲染引擎、卡片模板/风格库、xhs-post skill(流程编排 + 通用小红书规则)、MCP 发布接入
lab/(嵌套私有 repo redpress-lab,被 gitignore):真实帖子档案、互动数据、账号文风基线、经验笔记- 个人化的东西(账号偏好、数据结论)只进 lab,不进本仓库——工具层保持任何人可用
核心产品原则
- 改写力度是用户的显式选择,不是默认行为。收到草稿后必须先问力度档位(见
.claude/skills/xhs-post/SKILL.md的 L0-L4),不要自作主张大改 - 文案两级机制:本仓库 skill 写通用规则(平台限制、结构方法论),
lab/notes/写账号个性基线;账号基线覆盖通用默认,写文案前先读 lab/notes - 图片风格走 DESIGN.md 沉淀:HTML 生图的色板/字体/构图规范逐渐累积到 DESIGN.md(待建,需要专门讨论),模板系统朝多风格演进(
templates/<style>/)
架构
编排层 .claude/skills/xhs-post/SKILL.md 流程 + 通用文案规则
渲染层 scripts/render.ts + templates/ post.json → 3:4 PNG 卡片
发布层 两个并列产出末端:
自动 xiaohongshu-mcp(外部) localhost:18060/mcp
⚠️ vendor/ 二进制是自维护 fork(jackjin1997/xiaohongshu-mcp)
= upstream + #592(评论少时找不到评论)+ #711(导航重试,我们提的)
上游修复合并前不要用官方 release 覆盖;fork 源码在 ../xiaohongshu-mcp-fork
人工 scripts/handoff.ts + src/handoff.ts cards/ → 自包含 handoff.html
文字一键复制、图片下载/复制,手动发;唯一零风控敞口的发布路径
数据层 lab/scripts/track.ts 互动快照 → lab/metrics/
post.json 契约见 src/types.ts(标题 ≤20 字、正文 ≤1000 字、tags 不带 #、首卡 cover)。
开发约定
- TypeScript strict,运行时 Bun(见下方 Bun 指引)
- 测试
bun test;渲染改动跑bun scripts/render.ts posts/fixture/目检 - conventional commits,不加 Co-Authored-By
- ⚠️
cookies.json(小红书登录态)和lab/绝不提交到本仓库
风控红线
- 同一账号不要在其他网页端登录(会踢掉 MCP 登录态)
- 发布前必须过用户确认门;发布失败先查登录态,不盲目重试
Bun 使用指引
Default to using Bun instead of Node.js.
- Use
bun <file>instead ofnode <file>orts-node <file> - Use
bun testinstead ofjestorvitest - Use
bun installinstead ofnpm installoryarn installorpnpm install - Use
bunx <package> <command>instead ofnpx <package> <command> - Bun automatically loads .env, so don't use dotenv.
- Prefer
Bun.fileovernode:fs's readFile/writeFile Bun.$\ls`` instead of execa.
更多 Bun API 参考 node_modules/bun-types/docs/**.mdx。
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.
- 3d ago First seen · 63 lines · 939 tokens per session scan A e6172374dce2
redpress CLAUDE.md is an instructions file published in the GitHub repository jackjin1997/redpress (2 stars, last pushed 1mo ago), licensed MIT. It adds 939 tokens to every session, about $0.0047 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 2026-08-31.
Other instructions, from other repositories
xiaohongshu-skill AGENTS.md
Instructions for DeliciousBuding/xiaohongshu-skill, covering agents.md, safety, execution, development and architecture.
redbook CLAUDE.md
Instructions for lucasygu/redbook, covering claude.md — redbook cli, publishing, build & test, project structure and render command (optional).
topology-gomoku AGENTS.md
Instructions for YinkaiYu/topology-gomoku, covering agents.md, 先确认工作位置, 身份与权限, 分支职责 and 平台专属适配.
social-media-toolkit AGENTS.md
Instructions for JNHFlow21/social-media-toolkit, covering agent development protocol, product boundary, invariants, secrets and change discipline.
content-agent CLAUDE.md
Instructions for qiuxchao/content-agent, covering content agent, 技术栈, 启动方式, api 服务 and 前端(另一个终端).
xiaohongshu-auto-ops-assistant AGENTS.md
AGENTS.md instructions for Sen-platotech/xiaohongshu-auto-ops-assistant, a project described as: 以人为最终责任节点的小红书内容生产、人工复核与受控发布工作流.