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 skills add KimYx0207/Kim_Service --skill xiaohongshu-skillgit clone --depth 1 https://github.com/KimYx0207/Kim_ServiceWrote 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/kimyx0207/kim_service/xiaohongshu-skill)<a href="https://agentmods.dev/skills/kimyx0207/kim_service/xiaohongshu-skill"><img src="https://agentmods.dev/badge/skills/kimyx0207/kim_service/xiaohongshu-skill/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/kimyx0207/kim_service/xiaohongshu-skill"><img src="https://agentmods.dev/badge/skills/kimyx0207/kim_service/xiaohongshu-skill.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 | $0.00104 | $0.04466 |
| Opus 5 | $0.00052 | $0.02233 |
| Sonnet 5 | $0.00021 | $0.00893 |
| Haiku 4.5 | $0.00010 | $0.00447 |
Grade A, and why
xiaohongshu-skill 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 13d 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 — 139 lines — stays where its author put it; the contents beside it link to each section on GitHub.
小红书图文技能
用于做一条能被用户直接验收的小红书/Rednote 图文笔记。默认先确认策略和视觉方向,再进入图片生成动作;不要把“图片提示词”当成图片产物,也不要一上来读取所有参考。完整 15 段工作台是归档和校验合同,不是默认聊天最终展示。
核心规则
- 先看真实材料、受众、边界和用户意图,不从爆款模板开始;随后自行研究读者会搜什么、在什么场景卡住、同类内容的常见切口和可验证的高信号样本。不要把“流量从哪里来”丢给用户回答。
- 原始用户输入是最高优先的任务事实;HookPrompt、优化后的完整提示词、模型自写 scope 或总结不得新增“只要文案”“不需要制作图片”“不出图”等用户没说过的限制。若原始输入和优化提示词冲突,以原始输入为准。
- 本 skill 没有纯文案捷径:一旦触发本 skill,就必须走策略、标题/封面字、视觉路线、封面 MVP、确认后批量生成和发布前自检流程。
- 策略路线、标题/封面字、视觉路线、图片 Brief、封面候选和批量出图许可,都属于用户决策点;模糊输入必须优先通过 Codex
request_user_input、Claude CodeAskUserQuestion或宿主等价原生弹框补全。 - 封面图和封面文字的 MVP 被用户确认后,即视为完整图文包的生产许可:必须继续生成至少 8 个标题、1 篇正文、封面和 6-8 张内页图;不得再把范围降级成“只写文案/只做脚本/只做封面”。
- 图片路线优先级不可跳级:Image2 / 宿主原生生图 > 其他图片生成 MCP > 已有素材 > SVG/HTML/本地排版。Image2 或其他图片 MCP 任一可用时,禁止使用 SVG/HTML/本地排版;两者都不可用或都失败时才允许降级,并写明降级证据和原因。权威定义见
references/skill-contract.md「图片路线合同」。 - 可选视觉工作台只负责承载已确定的 3:4 页面槽位、插入本轮生成资产、缩略图检查、批注返修对照和逐页清单;它不是新的生图层级,也不能在 Image2 失败后冒充替代生图能力。通用合同见
references/visual-workspace-contract.md,具体供应商只在适配清单中映射。 - 可选视觉工作台不在 Skill 启动时询问。只有机会研究、营销决策、逐页视觉方向和 Image2 Brief 已完成,且首张 3:4 封面候选已经真实生成、用户马上能比较排版或批注时,才询问一次是否打开;跳过、不可用或启动失败都继续现有封面验收和图片生成链。
- SVG、HTML 卡片和静态结构图只能算结构预览,不能冒充真实图片,不能写成交付级。
- 不得把推荐方案写成“用户已选择”。
human_confirmed只能来自用户原话或宿主原生选择工具返回;否则写blocked_pending_human_decision或explicit_auto_permission。 - 旧的本地输出目录、旧图片和历史线程结果只能作参考或对比,不能算“本次已生成”。除非用户明确要求复用旧图,否则生成类请求必须有本轮图片工具证据,或诚实写
blocked: no_live_image_generation_evidence。 - 宿主原生生图可作为 Image2 等价主路径;证据可以是
imageGeneration、image_gen、本轮新生成图片路径或宿主可见图片结果。不要在宿主能生图时跳到 MCP,也不要把“当前无名为 Image2 的工具”写成跳过原生生图的理由。 - 课程、咨询、服务、知识产品相关笔记,公开区避免“私信我、加群、报名、领资料、外链、联系方式”等导流话术。
- 最终交付必须在聊天里可见,但必须是发布版摘要:图片预览/图片清单、选中标题、标题备选、正文、承接边界、标签、缺少的真实材料。不要把 15 段工作台、策略过程、提示词全量倒进最终聊天,除非用户明确要完整工作台。
- 用户已经要求
图文笔记、完整发布包、引流贴/种草贴、封面/内页、图片生成或“帮我生成”时,默认目标就是完整流程。不要再把“文案+脚本、不生成图片”设为推荐,也不要在前置决策完成后追问“是否只要文案”。 - 用户说“帮我生成 / 做一组 / 适合发布 / 图文 / 照片或画面加文字”,且没有说“先等我确认”时,视为首轮封面 MVP 生成许可:状态写
explicit_auto_permission,来源写“原始生成请求触发首轮封面 MVP 许可”。AI 可以选择保守默认策略、标题、视觉路线和 Brief,并生成 1 张封面候选用于验收;不得把该许可当成批量内页许可。只有用户原话明确“直接全套 / 不用问 / 你定并完整做完 / 直接做完整图文包”,或封面验收返回“通过并批量生成”,才允许生成封面 + 6-8 页内页。 - 新窗口、子线程或无人值守验收场景里,不要把“需要用户确认”当成默认停止理由;只有宿主拒绝、图片工具失败、宿主缺少生图能力或用户明确要求停顿,才允许
blocked。缺真实素材只能触发“AI/生成素材声明”和保守 Brief,不能自动降级成只给图片建议。 真实女生/美女/真人写真/留学生活/交友等真实感表达不是停图理由,也不是必须改成“虚构身份”的理由。按用户要的真实感做 Image2/宿主原生图像;只避免具体联系方式、未成年人和违法欺诈元素。文字信息卡、图片上文字说明是主体、大字封面、知识卡不是跳过 Image2 的理由,也不是直通本地排版的理由。Image2 或其他图片 MCP 任一可用时禁止本地排版;只有中文错字、不可读、排版失败或用户要求可编辑时,才降级为本地排版 PNG,并写明 fallback 原因。- 用户完成策略、标题/封面字、视觉路线等前置选择后,下一步是执行已选生成路线;不得再弹“这次要做到哪一步”这类范围选择。封面验收弹框里用户选择“通过并批量生成”后,必须继续生成封面 + 6-8 页内页图,不能只交付文案、脚本或单张封面。
What ships with it
48 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.
- assets/marketing-decision-board.md 4.8 KB
- assets/rednote-note-workbench.html 8.6 KB
- assets/visual-workspace-adapters/cowart.json 5.1 KB
- CHANGELOG.md 765 B
- contracts/visual-workspace-session.schema.json 6.1 KB
- docs/images/alipay.jpg 163 KB
- docs/images/contact-qr.png 825 KB
- docs/images/wechat-pay.jpg 70 KB
- evals/trigger-eval.json 6.9 KB
- evals/visual-workspace-offer.eval.json 1.3 KB
- examples/chat-output-bad-hookprompt-text-only.md 226 B
- examples/chat-output-bad-local-text-card.md 408 B
- examples/chat-output-bad-real-material-stop.md 397 B
- examples/chat-output-bad-text-script-recommendation.md 296 B
- examples/chat-output-bad-workbench-dump.md 545 B
- examples/chat-output-good.md 2.0 KB
- examples/example-input.md 1.1 KB
- examples/example-output.md 22 KB
- examples/page-direction-regression.md 1.0 KB
- LICENSE 1.0 KB
- NOTICE 209 B
- README.md 5.4 KB
- references/domain-research-brief.md 2.5 KB
- references/evaluation-method.md 4.1 KB
- references/failure-modes.md 14 KB
- references/human-decision-first-workflow.md 10 KB
- references/image-generation-workflow.md 7.9 KB
- references/interactive-mvp-decision-system.md 9.7 KB
- references/marketing-decision-model.md 4.4 KB
- references/page-art-direction-system.md 8.4 KB
- references/platform-risk-boundaries.md 1.3 KB
- references/provider-adapters/cowart.md 7.0 KB
- references/skill-contract.md 7.9 KB
- references/visual-system.md 2.3 KB
- references/visual-workspace-contract.md 7.6 KB
- references/writing-and-cover-quality.md 3.4 KB
- scripts/build-rednote-final-replacement-pack.mjs 17 KB runs code
- scripts/build-rednote-image2-feed-test.mjs 6.4 KB runs code
- scripts/build-rednote-realshot-handoff.mjs 21 KB runs code
- scripts/build-rednote-visual-pack.mjs 15 KB runs code
- scripts/check_acceptance_runs.py 6.0 KB runs code
- scripts/check-chat-output-fixtures.mjs 4.8 KB runs code
- scripts/check-narrative-route-fixtures.mjs 4.8 KB runs code
- scripts/check-page-direction-fixtures.mjs 9.5 KB runs code
- scripts/check-rednote-chat-output.mjs 5.5 KB runs code
- scripts/check-rednote-note.mjs 35 KB runs code
- scripts/check-visual-workspace-contract.mjs 22 KB runs code
- worksheets/real-detail-intake.md 1.2 KB
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.
- 13d ago First seen · 139 lines · 104 tokens per session scan A 3e6d0ce2718a
xiaohongshu-skill is a skill published in the GitHub repository KimYx0207/Kim_Service (169 stars, last pushed 1mo ago), licensed MIT. It adds 104 tokens to every session and 4,466 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 2026-08-30.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
insight-error-page
Write or audit an insight-kind error page for the Next.js dev overlay. Use when creating a new errors/ .mdx page, auditing an existing one, or checking that a page matches the framework fix cards. Covers page structure, title alignment, FixCard cards with Copy prompt button, code snippets, terminology verification…