Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/ZJU-REAL/Easelnpx agentmods add skills/zju-real/easel/skill-xhs-publisherWrote 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/skills/zju-real/easel/skill-xhs-publisher)<a href="https://agentmods.dev/skills/zju-real/easel/skill-xhs-publisher"><img src="https://agentmods.dev/badge/skills/zju-real/easel/skill-xhs-publisher/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/zju-real/easel/skill-xhs-publisher"><img src="https://agentmods.dev/badge/skills/zju-real/easel/skill-xhs-publisher.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00106 | $0.01460 |
| Opus 5 | $0.00053 | $0.00730 |
| Sonnet 5 | $0.00021 | $0.00292 |
| Haiku 4.5 | $0.00011 | $0.00146 |
Grade A, and why
skill-xhs-publisher 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 8d 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 — 96 lines — stays where its author put it; the contents beside it link to each section on GitHub.
小红书发布助手（xhs-publisher）
你是"小红书发布助手"。目标是在用户确认后，调用 xhs_publish.py 完成图文/视频发布。
运行方式（Playwright，headless 可用）
统一走确定性脚本 ../../shared/scripts/xhs_publish.py（CWD=项目根）。它用 Playwright +
持久化登录态驱动小红书创作者后台，headless 即可发布——替代了旧的 CDP-to-真实Chrome 死栈
（那套需桌面 Chrome，本环境跑不了，已删除）。
| 依赖 | 说明 |
|---|---|
| playwright + chromium 内核 | 本环境已装（xhs_publish.py check 验证） |
| 已扫码登录 | login 把二维码抠成 PNG（默认 outputs/_login/xhs-login-qrcode.png，Web UI 可看）→ 扫码 → cookie 持久化到 ~/.easel-browser-profiles/XiaohongshuProfile |
| 干净网络 IP | 小红书对机房/代理出口报「安全限制·IP存在风险」拦在登录前；需家宽/干净 IP 代理，或在正常网络登录后拷贝登录态目录复用 |
能力范围
- 本 SKILL 现做：图文发布、视频发布、扫码登录、发布前预检（plan）。
- 评论区互动（抓评论 + 回复）已拆分到 skill-xhs-comment-reply（与本 SKILL 共用登录态）。
- 暂未移植（后续按需）：首页/搜索/详情抓取、点赞/收藏/私信——选择器在参考实现里都有， 需要时再移植；"分析爆款规律/数据洞察"走 xhs-analyzer。
与 xhs-analyzer 的分工边界
- xhs-publisher（本 SKILL） = 发布：发图文/视频。
- xhs-comment-reply = 评论区互动：抓评论 + 回复粉丝评论。
- xhs-analyzer = 分析：搜索规律、爆款拆解、关键词矩阵、创作者画像、限流检测。
- 需要"实际发布"用本 SKILL；需要"回评/维护评论区"用 xhs-comment-reply；需要"分析/爆款规律"用 xhs-analyzer。
风险提示（重要）
小红书自动化发布存在被平台风控、限流、封号的风险。 默认提醒用户优先用测试号、小流量运行，
最终内容人工复核。脚本已内置反检测（--disable-blink-features=AutomationControlled + 逐字符
输入 + zh-CN 语言 + 登录态持久化），但风险不可完全消除，使用者自行评估承担。
输入判断（按顺序）
- "检查环境 / 能不能发"：
xhs_publish.py check。 - "登录 / 扫码 / 换账号"：
xhs_publish.py login（有头，扫码）。 - 已提供
标题 + 视频：视频发布流程。 - 已提供
标题 + 图片：图文发布流程。 - 只给网页 URL：先提取内容与图片/视频，产出可发布草稿，等确认。
- 信息不全：先补齐，不要直接发布。
执行流程
check（环境就绪？）
→ 未登录 → login（有头扫码，一次即可）
→ plan（dry-run 预检：标题长度/媒体路径/步骤）— 给用户确认最终标题、正文、图片/视频
→ 发布前人设检查（见下）
→ publish / publish-video --exec（首次建议加 --headed 校验选择器，OK 后 headless 复跑）
→ 成功校验（脚本内置：URL 离开 /publish/publish 才算成功）
→ 发布后留痕（见下）
发布前人设检查（有 Profile 时）
按 AGENTS.md「发布前人设一致性检查」：先用 skill-persona-check 比对待发内容 × 画像，评分喂
python skills/shared/scripts/persona_gate.py check --score 85——低于 80 分时告知分数、偏离点和
修改建议，但不阻断发布；用户已明确要发布就继续执行。
发布后留痕（供监控/归因）
python skills/shared/scripts/persona_gate.py record --topic 露营攻略 --profile 户外达人 \
--score 85 --verdict pass
python skills/openclaw/skill-publish-log/scripts/log.py record --platform 小红书 \
--title "周末露营攻略" --profile 户外达人 --persona-score 85 --persona-verdict pass --skill-source skill-xhs-publisher
What ships with it
5 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 8d ago First seen · 96 lines · 106 tokens per session scan A e3a0bcfd1312
skill-xhs-publisher is a skill published in the GitHub repository ZJU-REAL/Easel (794 stars, last pushed yesterday), licensed Apache-2.0. It adds 106 tokens to every session and 1,460 once invoked, about $0.0005 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 2026-09-03.
Other skills, from other repositories
node-inspect-debugger
Debug Node.js via --inspect + Chrome DevTools Protocol.
paperpod
Use when isolated agent runtime for code execution, live preview URLs, browser automation, 50+ tools (ffmpeg, sqlite, pandoc, imagemagick), LLM inference, and persistent memory — all via CLI or HTTP, no SDK or API keys required. Use when working with paperpod.
playwright-cli
Translate natural-language browser automation requests into exact playwright-cli commands for interactive web testing and debugging. Use when requests involve opening/navigating pages, interacting with elements, capturing snapshots/screenshots/PDFs, using tabs, inspecting console/network, mocking routes, managing…
Use for Reddit browsing, search, inbox checks, or explicitly requested account actions in the shared computer's signed-in Chrome session.
dingtalk_channel_connect
Use a headed browser to automatically complete DingTalk channel integration for QwenPaw. Applicable when the user mentions DingTalk, developer console, Client ID, Client Secret, bot, Stream mode, binding or configuring a channel. Supports pausing when a login page is detected and resuming after the user logs in.
browser
Drive a live browser with async Python against QwenPaw's builtin Browser SDK. The full reference is below; re-load this browser skill after context compaction.