技能详情（站内镜像，无评论）
许可证：MIT-0
MIT-0 ·免费使用、修改和重新分发。无需归因。
版本：v0.1.0
统计：⭐ 0 · 528 · 3 current installs · 5 all-time installs
⭐ 0
安装量（当前） 5
🛡 VirusTotal ：可疑 · OpenClaw ：可疑
Package：a950701zz/xiaohongshu-assistant-operator
安全扫描（ClawHub）
- VirusTotal ：可疑
- OpenClaw ：可疑
OpenClaw 评估
The skill's described behavior (automated reading, publishing and validating posts for a specific Xiaohongshu account) is plausible, but the runtime instructions ask the agent to perform privileged actions (read notes, publish posts, confirm via management pages) while the package declares no credentials, no install, and references a missing helper script — these gaps make the skill internally inconsistent and require clarification before use.
目的
The skill's name and description match the instructions (operating a single creator account end-to-end). However, the workflow requires account-level access (reading creator notes, submitting posts, checking the creator management page) but the skill declares no credentials, auth method, or config paths. That mismatch (wanting to perform privileged platform actions while requesting no auth) is incoherent.
说明范围
SKILL.md explicitly instructs the agent to read the last 5 creator notes, submit posts (image-text mode), navigate the creator management page to validate posts, and run scheduled comment interactions. These are concrete I/O actions against an external platform but the instructions do not specify how to authenticate, what endpoints or UI automation to use, nor do they provide or include the referenced publish_strict.py script. The open-ended '…
安装机制
There is no install spec and no code files beyond plain text references. That minimizes install-time risk, but also means the skill depends on the agent's existing environment (browsers, sessions, or external connectors) which is not described.
证书
The skill performs actions that normally require authentication tokens or browser sessions (publish, management-list checks, reading private creator notes) but declares no required environment variables or credentials. This is disproportionate: either the skill assumes implicit access to the user's Xiaohongshu account (not declared) or the SKILL.md is incomplete. Either case is a red flag.
持久
always is false (normal) and autonomous invocation is allowed (platform default). Autonomous operation plus the ability to publish content increases potential impact if credentials are later supplied, but autonomy alone is not unusual — the main concern is the missing authentication/operation details combined with autonomous publishing capability.
安装（复制给龙虾 AI）
将下方整段复制到龙虾中文库对话中，由龙虾按 SKILL.md 完成安装。
请把本段交给龙虾中文库（龙虾 AI）执行：为本机安装 OpenClaw 技能「Xiaohongshu Assistant Operator」。简介：Automates daily Xiaohongshu content creation, publishing, marketing, and engage…。
请 fetch 以下地址读取 SKILL.md 并按文档完成安装：https://raw.githubusercontent.com/openclaw/skills/refs/heads/main/skills/a950701zz/xiaohongshu-assistant-operator/SKILL.md
（来源：yingzhi8.cn 技能库）SKILL.md
---
name: xiaohongshu-assistant-operator
version: 0.1.0
description: Fully automated Xiaohongshu content operator for the specific creator "Bu Zhuan Dao Da Mo Xing Bu Gai Ming" (ID: 4740535877). Use when generating, publishing, and managing daily Xiaohongshu notes as the creator's assistant, including: (1) reading recent notes to infer direction, (2) scanning industry trends (LLM / Agent / RAG), (3) deciding daily themes (3 posts/day), (4) controlled promotion of the 199 note only, (5) cold-start strategy management, (6) automated publishing with strict validation, and (7) scheduled comment interaction. Trigger when operating this specific account end-to-end.
---
# Xiaohongshu Assistant Operator
This skill operates as the exclusive AI assistant for the Xiaohongshu creator:
> Bu Zhuan Dao Da Mo Xing Bu Gai Ming
> Xiaohongshu ID: 4740535877
It provides a fully automated daily content and interaction system including topic discovery, post generation, controlled marketing, publishing validation, and comment engagement.
This skill is hard-coded for this creator and must not be reused for other accounts.
---
# Core Workflow
Follow this high-level workflow. Load detailed logic from `references/` when executing.
1. Read last 5 creator notes.
2. Detect recent themes and 199 exposure frequency.
3. Scan industry trends (LLM / Agent / RAG / AI job market).
4. Determine operating phase (cold start or stable).
5. Generate 3 posts (<= 400 characters, image-text mode).
6. Enforce persona template (see `references/persona-template.md`).
7. Apply marketing intensity rules (see `references/marketing-control.md`).
8. Publish with strict validation (see `references/strict-publish-validation.md`).
9. Execute scheduled comment interaction (see `references/comment-reply-logic.md`).
---
# Operating Phases
Load detailed strategy from:
- `references/cold-start-strategy.md`
- `references/topic-decision-engine.md`
Phase detection rule:
- Days 1–7: Cold Start
- Day 8+: Stable Growth
---
# Hard Constraints
- Must always use assistant persona
- Must always @ the creator
- Only allowed promotion: 199 note
- Max 1 strong promotion per day
- Max 3 posts per day
- Max 3 interaction cycles per day
---
# Scheduling Rules
Default cadence:
- 3 posts: morning / afternoon / evening
- 3 interaction cycles evenly spaced
See `references/schedule-system.md` for timing logic.
---
# Manual Trigger Mode
When manually triggered:
- Execute one full daily cycle
- Publish 3 posts
- Run 1 interaction cycle
---
# Resource Map
## references/
- persona-template.md
- cold-start-strategy.md
- topic-decision-engine.md
- marketing-control.md
- dynamic-adjustment.md
- comment-reply-logic.md
- schedule-system.md
- risk-control.md
- strict-publish-validation.md
## scripts/
Optional deterministic utilities:
- publish_strict.py
---
This skill is a dedicated operator for one creator and must not be generalized.