Auto Contents
WarnAudited by ClawScan on May 10, 2026.
Overview
The skill is coherent but powerful: it can automatically send or publish AI-generated content through connected WeChat, Feishu, and Xiaohongshu accounts.
Install or invoke this only if you trust the local MakeContents service and are comfortable with agent-generated content being sent to configured channels. Prefer draft/test destinations, restrict provider credentials, keep Xiaohongshu publishing disabled unless needed, and review the persistent agent-rules.md file regularly.
Findings (4)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Incorrect, low-quality, or unintended AI-generated content could be sent to users or written into a knowledge base before a human reviews it.
The skill instructs the agent to generate and distribute content directly to connected channels. The artifacts do not show a mandatory human approval step before WeChat/Feishu distribution.
筛选推送 — 根据学习到的规律选择有价值的资讯，直接生成推送内容并分发到微信和飞书知识库
Use this only with trusted/test destinations or add a human approval gate before any WeChat or Feishu push. Limit the number of items and verify generated text before enabling unattended operation.
One bad selection or summary can spread across messaging, knowledge-base records, and local saved state.
A single agent-selected push action automatically propagates to multiple systems and changes saved state, with no rollback or containment described.
推送后自动： - 微信推送 - 飞书知识库写入 - 标记资讯为已保存，记录 push_type
Stage outputs as drafts first, use a review queue, and keep an audit/rollback process for pushed or archived items.
The agent may act with the authority of the configured Feishu, WeChat, or Xiaohongshu accounts.
The skill depends on provider API keys and a Xiaohongshu session cookie held by the local app to act on external accounts; the artifacts do not clearly bound account scopes or permissions.
各功能所需的 API Key（飞书、微信）已在应用配置页填写 ... 小红书 Cookie 已配置 + 发布开关已开启
Use least-privilege provider credentials, confirm the exact accounts and destinations configured in MakeContents, and keep Xiaohongshu publishing disabled unless intentionally needed.
Old, mistaken, or manipulated preference rules could affect what the agent chooses to publish later.
The skill creates persistent learned preference rules that influence future content selection and are retained over time.
此文件由 Agent 在执行"学习规律"流程后自动维护...每次执行学习流程后追加或更新对应章节，不删除历史记录。
Periodically review and edit or clear references/agent-rules.md, especially before enabling unattended publishing.