dbs-content-risk-check Skill (dbskill) for Claude: What It Does and How to Install It

Checks a post, script or profile before you publish, and separates what platform filters may flag from real content problems.

Installs: 10.7K; on skills.sh
Stars: 11K; on GitHub
Files: 1; besides SKILL.md
Updated: Oct 2026; last repo commit

This skill reviews titles, body text, image text, subtitles, voiceover, comments, profiles and video frames before you publish. It reports two things separately: what platform machine review may pick up (links, contact details, QR codes, ad structure, word lists), and what is actually wrong with the content (fraud, false claims, restricted promotion, medical claims, privacy and similar).

For each point it gives the exact location and the smallest change, and it keeps strong opinions and personal style. It can add a three-line caption for video context. SKILL.md is written in Chinese, gives no "can publish" verdict and no legal advice, and does not write wording that evades review.

Good for

Creators posting on Chinese social platforms such as Xiaohongshu
Editors and social managers proofing content before it goes out
Marketers checking ad claims, disclosures and contact details
People writing hiring or service posts

Install dbs-content-risk-check

Claude Code

Copy this prompt.

Install the dbs-content-risk-check skill in Claude Code. Run this command:

claude plugin marketplace add dontbesilent2025/dbskill
claude plugin install dbs-content-risk-check@dontbesilent-skills

Then check that it loaded and tell me how to use it.

2
Paste it into Claude Code. It shows the change and asks before running it.

Pick one route. The plugin route installs only this skill (the marketplace file points the dbs-content-risk-check plugin at this skill's folder, 2 files, and the repo tree has no hook files). The README's own skills CLI command installs the whole dbskill set to every agent it detects, so the box shows a single-skill command instead. With the plugin route, commands carry a dbs: prefix, for example /dbs:dbs-content-risk-check.

Built from the package data we hold. Not yet tested by us. Source.

What's in the skill

When it loads

面向创作者的内容发布风险检查。逐句检查自媒体、社交媒体和内容平台的标题、正文、图片文字、字幕、口播、评论、账号资料及视频画面，分别判断平台机器可能识别什么、内容本身存在什么实质问题，并给出具体位置和最小修改动作；需要明确视频语境边界时，额外生成真实准确的三行贴片小字。用户询问「检查敏感词」「发布前排雷」「这篇内容哪里可能违规」「这条小红书能不能发」「视频上放什么声明」时使用。保留强观点、情绪张力、个人风格和传播钩子。

Sections in SKILL.md

核心原则
工作任务
审核流程
第一步：识别材料和发布情境
第二步：检查五个表面
第三步：判断平台机器审核风险
第四步：判断内容实质风险
第五步：给出最小修改动作
最小必要干预
高频风险机制
内容安全与公共风险
网络访问与海外账号
玄学占卜与封建迷信
轻度性暗示、缩写和变体辱骂
受限推广与全局高风险主题
商业广告、招聘和接单

Key facts

Author: dontbesilent (maker per the README)
License: CC BY-NC 4.0 per the README badge and LICENSE file (the crawl reports NOASSERTION). Commercial use needs the author's permission
Installs on skills.sh: 10,732 (Nov 2, 2026)
GitHub stars: 11,039 for the whole dontbesilent2025/dbskill repo (Nov 2, 2026)
Last repo commit: Oct 28, 2026
What it includes: 2 files including SKILL.md: agents/openai.yaml (no scripts)
Needs: No software. It looks up a platform's current official rules only when the answer depends on them
Part of: dbskill 3.1.0, which the README describes as 34 Skills. The marketplace file lists this skill as its own plugin
Works in: Doubao, WorkBuddy, Claude Code, Codex and other agents that support Skills, per the README

What it includes

Platform, content location, purpose and who benefits
Works with partial material and says what was not checked, for example when only text was supplied.
Read · Identify material and context
Text, full meaning, visuals and sound, relationships, platform rules
A table of what to check on each surface, from subtitles and voiceover to resumes, QR codes and faces.
Try: “Check this Xiaohongshu draft for anything that could get it limited before I post.”
Read · Check five surfaces
Links, contact details, QR codes, ad structure, list-type words
Labelled as "may trigger machine review" or "may hit a list". It states no percentages and no promised outcomes.
Read · Machine review risks
High risk, needs confirmation, low-confidence leads
Needs a named object plus an action, result or purpose. Flags fraud, false claims, restricted promotion, medical claims, infringement, privacy and attacks.
Read · Content problems
Delete, add a condition, add proof, report, or use an official channel
Offers a safer version and a stronger version when a risky phrase carries the main selling point.
Write · Smallest fix
Real, accurate on-screen text for video context
Only text, no images. For example three short lines saying it is personal experience, not a guarantee, and to consider your own situation.
Write · Three-line video caption
Official rules only when the answer depends on them
Prefers official community rules, creator centers and ad platforms, and records the rule title and date.
Network · Platform rules lookup
Conclusion, priority fixes, machine flags, things to confirm
A compact report. It never says "can publish" or "cannot publish".
Read · Report format

From the official docs

Permissions

Low risk

What it can do

Read your data
Create and change things
Delete things
Read-only mode

Watch out for

It opens links only when you ask
SKILL.md treats a link in your draft as text to review and does not open it by default. It opens a page only when you ask to check the page, landing page or video, and it never opens or scans short links or QR codes on its own. github.com
The plugin and CLI routes differ in scope
The plugin route installs this one skill. The README's CLI command installs the whole set to every detected agent, so use the single-skill command if you want just this one. github.com

Security advisories

No published security advisories as of Nov 2, 2026.

Technical details

Turning on read-only	No read-only switch. The skill bundles no scripts.
Sign-in and keys	SKILL.md mentions no account or key.
Scopes	Claude reads the drafts, images, video or profile text you provide Claude may look up a platform's current official rules when the answer depends on them

Questions

What does the dbs-content-risk-check skill do?

It checks content line by line before you publish it. Claude reports what platform filters may flag and what is wrong with the content itself, then suggests the smallest fixes.

How do I install dbs-content-risk-check?

Use the install box at the top of this page. The plugin route installs only this skill, and the single-skill CLI command does too. The README's own CLI command installs the whole dbskill set to every agent it detects.

Will it tell me whether I can post?

No. SKILL.md says it does not give a can-publish or cannot-publish verdict, does not promise how a platform will react, and does not give legal advice.

Will it help me get around review?

No. SKILL.md forbids homophones, split characters, hidden contact details, comment codes and other wording meant to evade review.

Can I use it at work?

Check the license first. The repo is CC BY-NC 4.0 per the README and LICENSE, and the README says commercial use needs the author's permission.

What is the /dbs command?

A router skill in the same dbskill set. SKILL.md says that when /dbs is installed, Claude can suggest it for requests outside this skill. This skill works on its own.

Is it safe?

It bundles no scripts. Claude reads the text, images or video you give it. SKILL.md says links in your draft are not opened unless you ask for the page to be checked, and short links and QR codes are not opened or scanned.

Information

updated Nov 2, 2026

Repository: dontbesilent2025/dbskill
Folder: skills/dbs-content-risk-check

dontbesilent2025/dbskill skills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 4

dbs-content-risk-check SKILL.md · github.com
dbskill README · github.com
dbskill marketplace.json · github.com
dbskill LICENSE · github.com