dbs-resonate (dbskill): a Chinese-language skill that diagnoses why a draft misses its audience
A Chinese-language skill from the dbskill set that diagnoses where a finished draft fails to reach its target audience.
- Installs
- 19.3K
- on skills.sh
- Stars
- 11K
- on GitHub
- Files
- 1
- besides SKILL.md
- Updated
- Oct 2026
- last repo commit
dbs-resonate makes Claude list every claim in your draft, pick the single core claim and the action you want from readers, and choose one main communication theory to diagnose it. The theories are social identity, emotional appraisal, framing, uses and gratifications, and online sharing. It then names the sentence where the mechanism breaks and says what to delete, compress, strengthen and leave alone.
The skill is written and answers in Chinese. It does not predict views or click rates, promise that a draft will go viral, or write the revised text for you. It is one file plus a small agents/openai.yaml, and it runs no scripts.
Good for
- Chinese-language writers and creators with a draft that is not landing
- Editors who want sentence-level cut, compress and strengthen advice
- Content teams checking one core claim per article
Install dbs-resonate
Claude Code
- 1
Copy this prompt.
Install the dbs-resonate skill in Claude Code. Run this command: npx skills add https://github.com/dontbesilent2025/dbskill --skill dbs-resonate 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.
Requires Node.js.
From its listing on skills.sh. Source.
What's in the skill
When it loads
诊断已经写成的稿件哪里没有打中目标受众，结合原句解释共鸣断点并提出具体删改建议。适用于稿件修订；不负责整体流量与商业价值评估，也不以解释已发布内容的传播机制为主要交付。
Sections in SKILL.md
- 这个 skill 做什么
- 近邻任务边界
- 核心原则
- 1. 一篇文稿只保留一个核心主张
- 2. 一次诊断只选一个主理论
- 3. 诊断必须指向原文
- 4. 区分文稿能力与平台结果
- 可用理论与适用范围
- 条件理论
- 工作流程
- Step 1：提取全部主张
- Step 2：确定核心主张和目标动作
- Step 3：选择主导机制
- Step 4：定位问题
- Step 5：给出修改动作
- 输出格式
Key facts
- Installs on skills.sh
- 19,322 as of 2026-11-02
- Last repo commit
- 2026-10-28
- skills.sh audits
- Gen Agent Trust Hub: pass, Socket: pass, Snyk: pass
What it includes
List every explicit and implied claim before judging
Lists the draft's claims without evaluating them.
Try: “Use dbs-resonate on this draft and tell me which sentences lose the reader.”
Read ·
Step 1: extract all claimsOne core claim, one reader action
Chooses the claim to keep and the main reader action: understand, agree, relief, save, comment, share or act. If unclear it marks the core as unclear.
Read ·
Step 2: pick the core claim and target actionOne main theory, an extra only with clear evidence
Screens identity, framing, emotional appraisal, content use and sharing. Conditional theories (spiral of silence, cognitive dissonance, two-step flow) need their trigger conditions.
Read ·
Step 3: choose the main mechanismRate the mechanism as effective, weak or ineffective
Points to the sentence where identity, attribution, appraisal path, content use or sharing reason is missing.
Read ·
Step 4: locate the problemDelete, compress, strengthen, keep, and a one-line skeleton
Concrete changes tied to original sentences.
Write ·
Step 5: give edit actionsA fixed diagnosis report template, written in Chinese
Sections for the core problem in one sentence, the core mechanism review, the theory diagnosis and the specific edits.
Read ·
Output formatFive patterns, each with a fix
The patterns are vague audience, missing causal bridge, emotion labels only, unclear use and weak sharing reason.
Read ·
Common problem patternsSeven checks before finishing
Includes pointing every judgment to the original text and separating draft mechanism from platform results.
Read ·
Self-check
From the official docs
Questions
What does dbs-resonate do?
It diagnoses a finished or half-finished draft: what the core claim is, which communication mechanism is at work, where the text fails to deliver it, and what to cut, compress or strengthen. SKILL.md and its report template are written in Chinese; the file does not set a reply language, so check how it answers an English draft.
How do I install it?
Use the install box at the top of this page for this skill alone. The dbskill README also documents installing all 34 skills with one npx skills command and a Claude Code plugin route.
Does it write the revised draft?
No. Its handoff note lists the revised full text as something it does not deliver. It gives sentence-level edit advice and a one-line skeleton for the revised structure.
Can it predict views or whether a post will go viral?
No. SKILL.md lists predicting completion rate, click rate and recommendation volume as out of scope, and it separates what a draft can do from platform results.
Is it safe to use?
It works on the text you give it, and the folder holds SKILL.md plus agents/openai.yaml (a display name and default prompt), with no scripts. Its frontmatter lists no tools or API keys. Remember the license is CC BY-NC 4.0, so commercial use needs permission from the author.
How is it different from dbs-spread?
dbs-resonate works on a draft and tells you what to change, sentence by sentence. dbs-spread, another dbskill skill, explains why content that already got a reaction resonated or was shared, and does not edit the text. See dbs-spread's SKILL.md for its boundaries.
Information
updated Nov 2, 2026
- Repository
- dontbesilent2025/dbskill
- Folder
skills/dbs-resonate
Sources
Docs, issue trackers and discussions this page is built from.